Biosensors and Bioelectronics 19 (2004) 875883

Gold nanoparticle-based detection of genomic DNA targets on

microarrays using a novel optical detection system
James J. Storhoff , Sudhakar S. Marla, Paul Bao, Susan Hagenow, Hitesh Mehta,
Adam Lucas, Viswanadham Garimella, Tim Patno, Wes Buckingham,
William Cork, Uwe R. Mller
Nanosphere, Inc., Applied Research, 4088 Commercial Ave., Northbrook, IL 60062, USA
Received 17 February 2003; received in revised form 30 July 2003; accepted 22 August 2003

Abstract
The development of a nanoparticle-based detection methodology for sensitive and specific DNA-based diagnostic applications is described.
The technology utilizes gold nanoparticles derivatized with thiol modified oligonucleotides that are designed to bind complementary DNA
targets. A glass surface with arrays of immobilized oligonucleotide capture sequences is used to capture DNA targets, which are then detected
via hybridization to the gold nanoparticle probes. Amplification with silver allows for detection and quantitation by measuring evanescent
wave induced light scatter with low-cost optical detection systems. Compared to Cy3-based fluorescence, silver amplified gold nanoparticle
probes provide for a 1000-fold increase in sensitivity. Furthermore, direct detection of non-amplified genomic DNA from infectious agents
is afforded through increased specificity and even identification of single nucleotide polymorphisms (SNP) in human genomic DNA appears
feasible.
2003 Elsevier B.V. All rights reserved.
Keywords: Nanoparticles; DNA microarray; Silver amplification; SNPs; Staphylococcus; Thrombosis

1. Introduction
There has been an intense effort devoted toward development of new labeling and detection methodologies that
enable sensitive and low-cost detection of nucleic acids for
gene expression analysis and single nucleotide polymorphisms (SNP) identification (Lockhart and Winzeler, 2000;
Kwok, 2001; Niemeyer, 2001; Kirk et al., 2002). Current
and potential applications of such technologies include infectious disease detection, genetic disease predisposition or
diagnosis, and pharmacogenomics (Belgrader et al., 1998;
Lockhart and Winzeler, 2000; Kirk et al., 2002). Fluorescence technology has been the gold standard for detection
of DNA in both homogeneous reactions and on microarrays
due to its high sensitivity, dynamic range, and multiplexing capabilities (Duggan et al., 1999; Lipshutz et al., 1999;
Kwok, 2001). Despite these attributes, fluorescence labeling typically requires target amplification to obtain sufficient

Corresponding author. Fax: +1-847-562-88-86.

E-mail addresses: jstorhoff@nanosphere.us (J.J. Storhoff),
umuller@nanosphere.us (U.R. Mller).
0956-5663/$ see front matter 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.bios.2003.08.014

amounts of target and complex instrumentation is needed for

detection, ultimately limiting its utility in the applications
described above. There is a major need to develop more sensitive DNA detection methodologies at lower costs, with the
eventual goal of developing diagnostic systems that are simple and rapid enough to be used in a point of care setting.
One avenue that has been actively pursued over the past 8
years is the development of nanoparticle labeling methodologies (Storhoff and Mirkin, 1999; Niemeyer, 2001). Metal,
semiconductor, and magnetic nanoparticles offer a unique
set of physical properties that may be exploited in biological
detection assays (Mirkin et al., 1996; Bruchez et al., 1998;
Chan and Nie, 1998; Josephson et al., 2001). Of particular interest has been the use of metal nanoparticles, which
have been applied to the development of highly sensitive
nanoparticle-based detection assays that utilize electrical
or optical detection, including colorimetric and surface enhanced Raman spectroscopy (SERS) (Elghanian et al., 1997;
Yguerabide and Yguerabide, 1998; Taton et al., 2000, 2001;
Cao et al., 2002; Park et al., 2002; Walton et al., 2002).
Mirkin and coworkers were the first to report gold
nanoparticles as labels for DNA microarrays (Taton et al.,

876

J.J. Storhoff et al. / Biosensors and Bioelectronics 19 (2004) 875883

2. Methodologies
2.1. Materials

Fig. 1. Illustration of DNA hybridization to microarrays and detection

using silver amplified gold nanoparticle probes. Note that DNA target
and gold probe hybridization can be performed in a single step.

2000). The principle technology utilizes gold nanoparticles

derivatized with specific oligonucleotide sequences to detect DNA targets captured onto oligonucleotide microarrays
(Fig. 1). Although the nanoparticle labels can be detected
directly by absorbance when nanomolar (nM) concentrations of target are used in the hybridization to the glass
slide, a silver development methodology is employed to
amplify the signal associated with the particles by over five
orders of magnitude (Taton et al., 2000). In this method,
Ag(I) is reduced onto the gold nanoparticle surface with the
reducing agent hydroquinone, resulting in gold nanoparticle
induced growth of silver particulates. Notably, analysis of
the silver amplified gold nanoparticles with a simple flatbed
scanner provided sufficient sensitivity to detect a 50 fM
concentration of a 30-mer single-stranded DNA target in an
overnight hybridization assay (Taton et al., 2000).
In addition, the nanoparticle label/silver methodology
was demonstrated to be superior to a Cy3 labeled probe
for discriminating single nucleotide polymorphisms in a
sandwich assay. In a direct comparison, a G:T mismatch
pair could be differentiated from the correct A:T pair with
a 10:1 discrimination ratio with the nanoparticle labeling
system compared with only a 2.6:1 discrimination ratio
using a Cy3 fluorophore label. This enhanced discrimination is a result of the sharp melting transition conferred by
the nanoparticles onto the interacting DNA strands, while
fluorophore labeled DNAs maintain their much broader
transition profiles. This initial study provided the basis for
the development of simple, inexpensive nanoparticle-based
labeling and detection systems that provide sensitivity and
specificity superior to fluorescence. Herein, we describe the
application of waveguide technology and relatively simple
optics for the ultrasensitive detection of silver-amplified
gold probes in microarray-based assays. This labeling and
detection methodology provides for an 1000-fold increase
in detection sensitivity when compared to Cy3-based fluorescence. Significantly, we demonstrate that specific DNA
sequences within bacterial or human genomic DNA samples
are detectable without target amplification or complexity
reduction using this methodology.

HAuCl4 3H2 O, trisodium citrate, Tween 20, sodium nitrate, and Silver enhancer solution A and B were purchased
from SigmaAldrich. 20 SSC was obtained from Gibco
BRL, and Arrayit buffer and SuperAldehyde microarray
substrates were purchased from TeleChem International.
GAPSII slides were purchased from Corning Inc. PCR
reagents (10 PCR buffer, deoxynucleoside triphosphates,
and AmpliTaqGold polymerase) were purchased from
Perkin-Elmer. Oligonucleotide primers and amine modified oligonucleotides were ordered from Integrated DNA
Technologies or prepared in-house using an Expedite 8909
DNA synthesizer (Applied Biosystems). Silicone gaskets
(0.5 mm 9 mm) for microarrays were purchased from
Grace Biolabs. Methicillin resistant and Methicillin sensitive S. aureus genomic DNA samples (ATCC 700699 and
35556, respectively) were obtained from the American Type
Culture Collection (ATCC). Human placental DNA was obtained from SigmaAldrich and genotyped human genomic
DNA samples were obtained from Coriell Cell Repositories.
2.2. Gold probe preparation
Approximately 15 nm diameter gold nanoparticles were
prepared by the citrate reduction method (Grabar et al.,
1995). The approximate concentration of the gold nanoparticles was determined by measuring the particle size by
transmission electron microscopy (TEM) and gold atom
concentration by inductive coupled plasmaatomic emission spectroscopy (ICPAES). Optical spectra of the gold
nanoparticles were recorded with an HP8453 UV-Vis
spectrophotomer (max = 518 nm). The as prepared gold
nanoparticles were derivatized with thiol functionalized
oligonucleotides using a previously described salt aging
protocol (Storhoff et al., 1998). Briefly, the oligonucleotides
(4 M final concentration) are initially incubated with
the as prepared gold nanoparticles for >16 h, followed by
successive additions of phosphate buffered saline to a final concentration of 0.8 M NaCl, 10 mM phosphate (pH
7). After standing for >10 h, the probes were isolated by
centrifugation, washed in an equivalent amount of water, and then redispersed in 0.1 M PBS, 0.01% azide at a
particle concentration of 10 nM. All probes were stored
at 4 C.
2.3. Genomic DNA samples and PCR amplification
For the human genomic DNA hybridization experiments,
a 119 base-pair fragment of the MTHFR gene was PCR amplified from human placental DNA (wild type sequence confirmed by sequencing analysis) or fragmented by sonication
using a Misonix Ultrasonic Cell Disruptor Sonicator prior
to microarray hybridization. PCR amplification of human

J.J. Storhoff et al. / Biosensors and Bioelectronics 19 (2004) 875883

genomic DNA (100 ng) or Staphylococcus DNA (100 ng)

samples was carried out in 50 l volumes using 0.16 M
of primers and 2 mM deoxynucleotide triphosphates with
AmpliTaqGold polymerase on a GeneAmp PCR System
9700 (Perkin-Elmer). Following target denaturation at
95 C for 10 min, PCR was performed for 35 cycles, each
cycle consisting of 95 C for 30 s, annealing at 55 C for
30 s, extension at 72 C for 60 s, with a final extension step
at 72 C for 10 min. The PCR amplicon fragment size was
determined using a 1.0% Amplisize /Agarose gel (BIORAD) stained with ethidium bromide and visualized under
UV light. PCR products were purified using a GFXTM PCR
purification kit (Amersham Pharmacia Biotech) prior to
microarray hybridization.
2.4. Preparation of DNA microarrays, hybridization, and
image analysis
Amine modified oligonucleotide capture sequences
were arrayed onto SuperAldehyde slides following standard procedures recommended by TeleChem International
(http://www.arrayit.com) with an Affymetrix GMS 417
pin and ring microarrayer equipped with a 500 m diameter pin. Typically, 10 replicates of the capture sequences
were arrayed at defined locations across the glass slide so
that multiple tests could be performed on a single slide
by isolating reaction sites with silicone gaskets to create
individual wells. Microarray hybridization reactions were
carried out by applying 50 l of the probe/target mixtures
(typically 100 pM1 nM probe) in 5 SSC/0.05% Tween
20 to the microarray after heat denaturation at 95 C. Following incubation in a humidified chamber at the specified
temperature, slides were washed with 0.5 M NaNO3 , 0.05%
Tween 20 at room temperature for 1 min. The gaskets were
then removed and the slides were centrifuged to remove
excess liquid. For silver development, silver enhancer solutions A and B were mixed 1:1 and placed onto each slide
and incubated for 35 min. After incubation, slides were
washed with water and imaged using a commercial image
analysis system, or the NIS2000, a prototype Nanosphere
image analysis system. The commercial scanner was designed for scanning large areas at high resolution (Brown
et al., 2001). A metal halide arc lamp is used for illumination, which enables detection of fluorophores or silver
amplified metal nanoparticles with the appropriate interference filters. The fluorescence emission or scattering from
the illuminated slide is collected through a microscope
objective onto a cooled CCD camera. To image an entire
microarray, a series of smaller images (2.4 mm 2.4 mm)
are captured at high resolution by moving the slide under
the objective. The smaller images are then stitched together
for analysis of the entire array. The NIS2000 uses a 75 W
halogen bulb for evanescent illumination and its optical
train is designed to capture a standard microscope slide in
a single image. Typical image acquisition times are on the
order of 1 s.

877

Table 1
Sequence
mecA capture 1
mecA capture 2
mecA probe
Up primer
Down primer

5 -ATGGCATGAGTAACGAAGAATA-3
5 -TTCCAGATTACAACTTCACCA-3
5 -A15 -peg-GCACTTGTAAGCACACCTTCAT-3
5 -ATCCACCCTCAAACAGGTGA-3
5 -ACGTTGTAACCACCCCAAGA-3

Table 2
Sequence
5 -A20 -TAT TCC TCG CC-3

5 -A20 -ATT CCT TGC CT-3

5 -CTG CTC TTA CAG ATT AGA AGT

AGT CCT-3
5 -GAC ATC GCC TCT GGG CTA AT-3
5 -CTG AAA GGT TAC TTC AAG GAC-3

FV WT probe
FV Mut probe
FV capture
FV up primer
FV down primer
Table 3

Sequence
MTHFR probe
MTHFR capture
Negative control
MTHFR up primer
MTHFR down primer

5 -A20 -TGATGAAATCGGCT-3
5 -CCCGCAGACACCTTCTCCTTC-3
5 -CTGCTCTTACAGATTAGAAGTAGTCCT-3
5 -TATTGGCAGGTTACCCCAAA-3
5 -CTCACCTGGATGGGAAAGAT-3

2.5. Sequence tables

For sequence tables, see Tables 13.
3. Results and discussion
3.1. Comparison of nanoparticle labels and Cy3 labels
The first goal in the development of the gold nanoparticle
technology for microarray applications was to compare the
detection limit of the standard Cy3 fluorescence technology
with light scattering-based detection of silver amplified gold
particles, using commercially available slides, reagents and
instrumentation. In order to compare the number of nanoparticle probes detectable per unit area with the number of
Cy3 labels detectable per unit area, a serial dilution of Cy3
labeled oligonucleotides was microarrayed onto a GAPSII
slide, and the fluorescence from each set of spots was imaged and quantified using a commercial scanner (Fig. 2). The

Fig. 2. The detection limit of serially diluted Cy3 labeled oligonucleotides

printed onto GAPSII glass slides and imaged with a commercial scanner.

878

J.J. Storhoff et al. / Biosensors and Bioelectronics 19 (2004) 875883

Fig. 3. The detection limit of serially diluted nanoparicle probes printed onto glass slides and imaged using a commercial scanner. (A) Image of
nanoparticle probes after silver development. Inset shows enlarged image of microarrayed spot containing 200 gold nanoparticles. (B) Averaged net
intensity of all pixels in the spot area with a signal above background plotted with standard deviation of net signal intensity.

detection limit was determined to be 7.2 fluors/m2 , which

is close to the number determined for Cy3 using an Axon
Genepix microarray scanner at Corning Inc. (5 fluors/m2 ,
Dr. Paul Bao, personal communication). For comparison, a
serial dilution of 15 nm diameter gold nanoparticle probes
(functionalized with the same 32-mer sequences) was arrayed onto a GAPSII slide, followed by silver enhancement
for 5 min using Sigma silver solution. The Rayleigh scatter was measured using a commercial scanner. As shown
in Fig. 3A, individual silver amplified gold nanoparticles,
while diffraction limited, gave a sufficiently strong signal to
expose the CCD pixels and thereby became detectable by
this imaging method with a corresponding detection limit
of 0.001 gold probes/m2 (Fig. 3B). This set of experiments suggests that under typical laboratory conditions with
commercial instrumentation and reagents the silver amplified gold nanoparticle detection technology offers a >3 orders of magnitude sensitivity advantage when compared to
conventional Cy3 labeling, whereby we assume a single label moiety (fluorophor or nanoparticle) per detection probe.
Note that we assume that none of the nanoparticles were
washed off during the silver amplification step. In addition,
the average gold probe signal is quantifiable over three to
four orders of magnitude which is comparable to the dynamic range observed in fluorescence labeling.

3.2. Detection of the mecA gene in Methicillin resistant

Staphylococcus aureus
Given the potentially much higher sensitivity of
nanoparticle-based detection, we have applied this technology to the development of nucleic acid-based detection
assays for infectious disease targets. Because of the urgent
need for rapid analyses of nosocomial infections we have
focused on Staphylococcus speciation and identification
of antibiotic resistance genes common to Staphylococci
from cell culture. Fig. 4 shows a small array containing
two amine modified capture sequences (sequence Table 1)
specific for the mecA gene, which confers resistance to the
antibiotic methicillin. For detection, a thiol-modified probe
oligo complementary to a different portion of the mecA
gene was conjugated to 15 nm diameter gold probes. The
mecA probe and capture sequences were tested initially
by comparing signal intensities obtained with a positive
control (a 281 base-pair mecA gene PCR fragment) to that
of two negative controls (PCR fragments from the Tuf and
coagulase genes). All target sequences were amplified from
methicillin resistant S. aureus DNA. Probe/target hybridizations were performed in a single one hour hybridization
step at 40 C on the microarray slide followed by silver amplification. As shown in Fig. 4, a sequence specific signal

J.J. Storhoff et al. / Biosensors and Bioelectronics 19 (2004) 875883

879

Fig. 4. (A) Image of wells from microarray containing specified PCR amplicons. Fragment size shown in parentheses. Two different capture probes
(mecA1 and mecA2) were used. (B) Corresponding data analysis from commercial scanner. Net signal intensities plotted with standard deviation.

is generated only for the positive control sample containing

the mecA gene PCR fragment.
In order to test the sensitivity of this system a serial dilution of the mecA gene PCR amplicon was added as target
to separate reactions following the same hybridization protocol as above. Fig. 5 shows that 3 106 copies (100 fM,
50 l) of the mecA gene PCR amplicon were detectable in
a 1 hour hybridization. Although this sensitivity to target is
significantly below our theoretical sensitivity based on the
number of detectable probes (see Fig. 3), it was sufficient
for achieving our initial goal of detecting specific genes
from genomic DNA extracted from culture. It also should be
noted that the detection sensitivity achieved here is roughly
equivalent to the 50 fM sensitivity achieved by Mirkin and
coworkers with two important distinctions. First, the target and probe hybridization has been reduced to a single
one hour hybridization step from the previously reported
overnight target hybridization reaction. Second, monitoring
the scatter signal has enabled the silver amplification time
to be reduced from 15 to 5 min minimizing the amount of
background due to silver autonucleation.
The next experiment was designed to demonstrate the
feasibility of detecting the mecA gene in the presence of
genomic DNA without complexity reduction by PCR. Total
DNA extracted from cultures of methicillin resistant and methicillin sensitive S. aureus cells were sonicated to shear the
genomic DNA, followed by hybridization of serially diluted
DNA (250 ng1 g) to the microarray using the protocol de-

Fig. 5. Serial dilution of a PCR-amplified mecA gene sequence hybridized

to a microarray containing a mecA gene capture probe. The normalized
net signal intensity of three spots is averaged and plotted with the standard
deviation. Dashed line equals the average background from three control
spots plus three standard deviations around that value.

scribed above. Fig. 6 shows that only the hybridizations with

target DNA from methicillin resistant strains resulted in positive signals, demonstrating that the nanoparticle probe-based
detection methodology allows specific detection of chromosomal genes from bacterial DNA without the need for target
amplification and complexity reduction. Optimization of
the detection sensitivity and extension of the assay to speciation of coagulase negative Staphylococci are currently
underway.

Fig. 6. (A) Hybridization of serially diluted genomic DNA from either methicillin resistant S. aureus (MRSA) or methicillin sensitive S. Aureus (MSSA)
to a microarray containing three repeat spots for the mecA gene capture probe. (B) Corresponding data analysis using a commercial scanner.

880

J.J. Storhoff et al. / Biosensors and Bioelectronics 19 (2004) 875883

Fig. 7. NIS2000 image analyzer. (A) Picture of analyzer. (B) Images of post-silver amplification spots as captured by the analyzer (spot dimensions:
650 m diameter).

3.3. NIS2000 image analyzer

The NIS2000 image analyzer represents a low-cost alternative to currently available commercial imaging systems, and was specifically designed at Nanosphere to capture
evanescent wave induced light scatter from silver amplified
gold probes (Fig. 7). In the current configuration, the optics
of the NIS2000 projects a standard microscope slide onto a
CCD camera in a single image, requiring no moving parts
for image acquisition. The software provides for correction
of image distortions, spot recognition and quantification of
signals from silver amplified gold nanoparticles. For comparison to the commercial imager, the same slides shown in
Fig. 3 were imaged on the NIS2000. As shown in Fig. 8, a
detection limit of 0.0025 probes/m2 was achieved with
this system, corresponding to 400 probes in a 400 m diameter spot. This is remarkable considering the simplified
optics that are used for detection, and indicates that complex
instrumentation is not required to achieve extremely high
sensitivity using nanoparticle probe technology.

Taton et al., 2000). When taken together with the high sensitivity, this technology appears capable of achieving our goal.
Venous thrombosis causes significant morbidity and
mortality in the US, and may arise from a hereditary predisposition due to at least one mutation in any of some
12 genes (Kearon et al., 2000). We have selected several
SNPs that are of particular import due to their relatively
high occurrence (Lane et al., 1996a,b). A very frequent
mutation is the G A transition at position 1691 in the
Factor V gene termed Factor V Leiden (Bertina et al.,
1994). A sandwich assay consisting of a 27 base universal
capture sequence specific for the Factor V gene and a set of

3.4. Nanoparticle-based SNP detection

One of our goals at Nanosphere is the application of the
nanoparticle probe technology to the detection of SNPs in
human DNA without the need for PCR amplification. Given
the extreme complexity of the human genome at 1109 total
base pairs (note that the biological complexity is lower than
the total number of bases because of repeat sequences in the
human genome), this represents a formidable challenge for
oligonucleotide-based microarrays and so far has not been
possible without PCR or other target amplification methodologies. One reason is that the melting temperatures of short
oligos are too low to provide for the stringency needed for
SNP discrimination in the presence of total human DNA.
One of the principle advantages of nanoparticle probes are
the extremely sharp melting transitions and increased melting temperatures, which permit probe hybridization under
much more stringent conditions resulting in better discrimination of single base mismatches (Elghanian et al., 1997;

Fig. 8. The detection limit of nanoparticle probes analyzed using the

NIS2000 image analyzer. (A) Image of serially diluted nanoparticle probes
printed onto a glass slide and amplified with silver. (B) The net signal
intensity was averaged from three spots and is plotted with standard
deviations.

J.J. Storhoff et al. / Biosensors and Bioelectronics 19 (2004) 875883

Fig. 9. Illustrations of SNP-on-probe detection assay format. Replicate

wells are printed on a single glass slide to compare wild type and mutant
probe signals. A capture sequence located adjacent to the SNP region is
immobilized onto the glass slide for target binding.

nanoparticle probes containing either the wild type (WT)

or mutant (MUT) (11 base recognition element for each
probe) sequence was designed to identify the homozygous
wild type, homozygous mutant, and heterozygous (HET)
genotypes (Fig. 9, sequence Table 2). Six separate wells
containing replicate spots of the Factor V Leiden capture
probe and a silver amplification control sequence were arrayed on the same glass slide to compare the WT and MUT
probe signals for the three genotypes (Section 2.4). Human
genomic DNA samples from WT and Factor V Leiden mutants (homozygous and heterozygous) were initially PCR
amplified and used to test the genotyping capability of this
system. Probe/target solutions were hybridized to an array
containing the universal capture and a control sequence

881

for 30 min at 52 C followed by silver amplification. The

relative signal intensities obtained for the WT and MUT
probes were compared to determine the genotype of the
sample. As shown in Fig. 10, all three genotypes are easily
discriminated using the NIS2000 image analyzer on the
basis of signal intensity, with match/mismatch probe ratios
of over 30 for the WT and MUT targets, and the expected
1:1 ratio for the HET target. This discrimination power is
roughly six times better than could be achieved on average
with matched versus single base mismatched 25-mers in a
fluorescence-based microarray hybridization (Relogio et al.,
2002). This demonstrates the much increased discrimination power of nanoparticle probe technology compared to
standard oligonucleotide-based hybridization systems.
After demonstrating SNP discrimination capability with
PCR-amplified samples, we pursued the direct detection
of SNPs in human genomic DNA. As a first step towards
this goal, we designed a sandwich assay system to detect
the wild type sequence of the MTHFR gene (sequence
Table 3), the location for another SNP with relevance to
thrombosis. A capture sequence specific for the MTHFR
gene and a negative control capture sequence were arrayed
for testing. Samples of human genomic DNA (20 g) were
hybridized to the microarray with MTHFR nanoparticle
probe conjugates under stringent conditions in a single
one hour hybridization step, followed by silver amplification. An MTHFR gene PCR fragment served as a positive
control in a separate well, and a well with E. coli DNA
served as a negative control. The sonicated genomic DNA
sample is estimated to contain a total of 6 106 target
copies (200 fM, 50 l). As shown in Fig. 11, a strong signal
is observed for the MTHFR gene capture sequence when

Fig. 10. (A) NIS2000 image of a Factor V Leiden SNP genotyping assay. T refers to the target capture sequence, and PrC refers to a proprietary control.
Hom, homozygous; HET, heterozygous. (B) Data analysis using the NIS2000 image analyzer. Data averaged from three spots each.

Fig. 11. (A) NIS2000 image of the MTHFR gene sequence in human genomic DNA hybridized to a microarray with (PCR) and without (genomic DNA)
PCR amplification. (B) Data analysis from the NIS2000 image analyzer. Data averaged from three spots each.

882

J.J. Storhoff et al. / Biosensors and Bioelectronics 19 (2004) 875883

the unamplified genomic DNA sample is hybridized to the

array, while almost no signal is observed for the negative
control sequence or for the E. coli negative control. This
initial data suggests that the nanoparticle probe technology
may be used for direct detection of specific sequences in
human genomic DNA samples without the need for complexity reduction by PCR.

4. Conclusions
Our preliminary experiments have demonstrated that individual silver amplified nanoparticle probes can be detected
with a high-end commercially available detection system
with high resolution optics. The detection limit of 0.001
gold probes/m2 is roughly three to four orders of magnitude better than what this instrument can achieve with a Cy3
labeled oligonucleotide (7.2 fluorophores/m2 ) on glass
substrates that were optimized for fluorescence detection.
Even with much lower resolution optics silver amplified
gold probes can be detected down to 0.0025 particles/m2 ,
demonstrating the remarkable sensitivity of nanoparticle
probe technology, that allows for small and robust, yet
low-cost instrumentation.
In assays with genomic target DNAs, this has translated to
a detection limit of 6 106 copies of target DNA (200 fM,
50 l) in a 1 h hybridization to DNA microarrays. This technology appears very suitable for speciation of pathogenic
strains in nosocomial infections in general and specifically
for the rapid detection of methicillin resistance in S. aureus from cultured cells. More important are the implications for rapid and low-cost detection of human genetic
disease predispositions, since we have demonstrated that
the nanoparticle probe technology can detect specific sequences in unamplified human genomic DNA. Future work
will focus on further increasing sensitivity and discriminating various SNPs of relevance to human genetic disease
detection.
Acknowledgements
We gratefully acknowledge the National Institutes of
Health for supporting this work through SBIR awards 1
R43 GM62096-01A1 and 1 R43 HL65876-01. We also
acknowledge the support of Nanospheres oligonucleotide
synthesis, assay development, and engineering teams, as
well as Dr. Karen Kaul (Northwestern University) for helpful discussions.

References
Belgrader, P., Benett, W., Hadley, D., Long, G., Mariella, R., Milanovich,
F., Nasarabadi, S., Nelson, W., Richards, J., Stratton, P., 1998. Rapid
pathogen detection using a microchip PCR array Instrument. Clin.
Chem. 44, 21912194.

Bertina, R.M., Koeleman, B.P.C., Koster, T., Rosendaal, F.R., Dirven,

R.J., Deronde, H., Vandervelden, P.A., Reitsma, P.H., 1994. Mutation
in bloodcoagulation Factor-V associated with resistance to activated
protein-C. Nature 369, 6467.
Brown, C.S., Goodwin, P.C., Sorger, P.K., 2001. Image metrics in the
statistical analysis of DNA microarray data. Proc. Natl. Acad. Sci.
U.S.A. 98, 89448949.
Bruchez, M., Moronne, M., Gin, P., Weiss, S., Alivisatos, A.P., 1998.
Semiconductor nanocrystals as fluorescent biological labels. Science
281, 20132016.
Cao, Y.W., Jin, R., Mirkin, C.A., 2002. Nanoparticles with Raman spectroscopic fingerprints for DNA and RNA detection. Science 297,
15361540.
Chan, W.C.W., Nie, S.M., 1998. Quantum dot bioconjugates for ultrasensitive nonisotopic detection. Science 281, 20162018.
Duggan, D.J., Bittner, M., Chen, Y., Meltzer, P., Trent, J.M., 1999. Expression profiling using cDNA microarrays. Nat. Genet. Suppl. 21,
1014.
Elghanian, R., Storhoff, J.J., Mucic, R.C., Letsinger, R.L., Mirkin, C.A.,
1997. Selective colorimetric detection of polynucleotides based on the
distance-dependent optical properties of gold nanoparticles. Science
277, 10781081.
Grabar, K.C., Freeman, R.G., Hommer, M.B., Natan, M.J., 1995. Preparation and characterization of Au colloid monolayers. Anal. Chem.
67, 735743.
Josephson, L., Perez, J.M., Weissleder, R., 2001. Magnetic nanosensors
for the detection of oligonucleotide sequences. Angew. Chem. Int. Ed.
40, 32043206.
Kearon, C., Crowther, M., Hirsh, J., 2000. Management of patients with
hereditary hypercoagulable disorders. Annu. Rev. Med. 51, 169185.
Kirk, B.W., Feinsod, M., Favis, R., Kliman, R.M., Barany, F., 2002. Single
nucleotide polymorphism seeking long term association with complex
disease. Nucleic Acids Res. 30, 32953311.
Kwok, P.Y., 2001. Methods for genotyping single nucleotide polymorphisms. Annu. Rev. Genomics Hum. Genet. 2, 235258.
Lane, D.A., Mannucci, P.M., Bauer, K.A., Bertina, R.M., Bochkov, N.P.,
Boulyjenkov, V., Chandy, M., Dahlback, B., Ginter, E.K., Miletich,
J.P., Rosendaal, F.R., Seligsohn, U., 1996a. Inherited thrombophilia.
1. Thromb. Haemost. 76, 651662.
Lane, D.A., Mannucci, P.M., Bauer, K.A., Bertina, R.M., Bochkov, N.P.,
Boulyjenkov, V., Chandy, M., Dahlback, B., Ginter, E.K., Miletich,
J.P., Rosendaal, F.R., Seligsohn, U., 1996b. Inherited thrombophilia.
2. Thromb. Haemost. 76, 824834.
Lipshutz, R.J., Fodor, S.P.A., Gingeras, T.R., Lockhart, D.J., 1999. High
density synthetic oligonucleotide arrays. Nat. Genet. Suppl. 21, 2024.
Lockhart, D.J., Winzeler, E.A., 2000. Genomics, gene expression, and
DNA arrays. Nature 405, 827836.
Mirkin, C.A., Letsinger, R.L., Mucic, R.C., Storhoff, J.J., 1996. A
DNA-based method for rationally assembling nanoparticles into
macroscopic materials. Nature 382, 607609.
Niemeyer, C.M., 2001. Nanoparticles, proteins, and nucleic acids: biotechnology meets materials science. Angew. Chem. Int. Ed. 40, 4128
4158.
Park, S.J., Taton, T.A., Mirkin, C.A., 2002. Array-based electrical detection
of DNA with nanoparticle probes. Science 295, 15031506.
Relogio, A., Schwager, C., Richter, A., Ansorge, W., Valcarcel, J.,
2002. Optimization of oligonucleotide-based DNA microarrays. Nucleic Acids Res. 30, e51.
Storhoff, J.J., Mirkin, C.A., 1999. Programmed materials synthesis with
DNA. Chem. Rev. 99, 18491862.
Storhoff, J.J., Elghanian, R., Mucic, R.C., Mirkin, C.A., Letsinger, R.L.,
1998. One-pot colorimetric differentiation of polynucleotides with
single base imperfections using gold nanoparticle probes. J. Am.
Chem. Soc. 120, 19591964.
Taton, T.A., Mirkin, C.A., Letsinger, R.L., 2000. Scanometric DNA array
detection with nanoparticle probes. Science 289, 17571760.

J.J. Storhoff et al. / Biosensors and Bioelectronics 19 (2004) 875883

Taton, T.A., Lu, G., Mirkin, C.A., 2001. Two-color labeling of oligonucleotide arrays via size-selective scattering of nanoparticle probes. J.
Am. Chem. Soc. 123, 51645165.
Walton, I.D., Norton, S.M., Balasingham, A., He, L., Oviso Jr., D.F.,
Gupta, D., Raju, P.A., Natan, M.J., Freeman, R.G., 2002. Particles
for multiplexed analysis in solution: detection and identification of

883

striped metallic particles using optical microscopy. Anal. Chem. 74,

22402247.
Yguerabide, J., Yguerabide, E.E., 1998. Light-scattering submicroscopic
particles as highly fluorescent analogs and their use as tracer labels in
clinical and biological applications. II. Experimental characterization.
Anal. Biochem. 262, 157176.