| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
BioStar, Inc., 6655 Lookout Rd., Boulder, CO 80301.
a Author for correspondence. Fax 303-530-6627; e-mail rmo{at}csn.net
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: In this system, capture oligonucleotide probes are immobilized on the surface of the biosensor. Interaction of the capture probes with a complementary target and a biotinylated detector oligonucleotide allows initiation of formation of an organic thin film on the biosensor. Thin film formation is completed by enzymatic activity of peroxidase conjugated to an anti-biotin antibody. Peroxidase catalyzes deposition of an insoluble product onto the silicon surface, generating a uniform thin film. The increased thickness on the surface alters the perceived color of the biosensor through changes in the interference patterns of reflected light from the surface, causing a color change from gold to purple.
Results: The biosensor results may be evaluated by direct visual inspection or quantified by ellipsometry. Results are obtained in 25 min with a detection limit of 5 pmol/L (150 amol/sample). Selectivity of the biosensor is demonstrated by discrimination of single nucleotide mismatches. Multitarget arrays are also analyzed with the thin film biosensor, and the system is capable of detecting targets from human serum and urine.
Conclusions: The biosensor surface is inexpensive to produce, and the assay format is simple and rapid. The thin film biosensor is adaptable to a wide variety of nucleic acid detection applications, including rapid diagnostic testing for infectious disease panels, antibiotic resistance panels, or allelic discrimination of specific genetic markers.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The detection method used here generates a visual signal by the direct interaction of light with thin films formed on a modified silicon surface. The thin film is initiated by the binding of the target molecule to the optically coated surface and completed by a horseradish peroxidase (HRP)1 -catalyzed precipitation of insoluble product on the surface. The assay relies on a test surface that is antireflective to a specific wavelength (or range of wavelengths) of visible light. Destructive interference occurs when light of a specific wavelength is reflected from both the substrate/thin film interface and the air/thin film interface. Light of a specific wavelength reflected from the upper surface will be out of phase with respect to light of the same wavelength reflected from the substrate/thin film interface; thus, that wavelength will be attenuated by destructive interference. The attenuation of one or more wavelengths of visible light will produce an characteristic surface color. Any change in the optical thickness of the thin film, such as one caused by the deposition of target molecules on the surface, will produce a visible color change from gold to purple. Surfaces are constructed such that angstrom (Å) level changes in thickness produce color transitions to which the human eye is maximally sensitive. The method is sensitive to optical thickness changes as small as 10–20Å.
Qualitative interpretation of results is determined visually. Quantitative analysis of the results can be performed with the fixed polarizer ellipsometer (FPE) (7)(8). The FPE detects changes in the polarization state of linearly polarized light caused by reflection from thin films on the silicon surface. The polarization change is proportional to the thin film thickness and therefore relates to the concentration of analyte bound to the biosensor surface.
We report the application of thin film detection to construct a simple, highly sensitive and selective biosensor for detection of nucleic acid targets. The system is compatible with typical clinical matrices, and results are obtained in less than 30 min.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
oligonucleotide sequences
All oligonucleotides were synthesized by IDT and purified by HPLC.
The sequences used in this study are summarized in Table 1
.
|
assay format
The mecA capture probe (1 µmol/L) was spotted on the
surface in 0.1 mol/L carbonate buffer, pH 10.0, and dried overnight at
room temperature. To begin the assay, the surface was rinsed with
distilled H2O to remove unbound capture
probe. mecA target (30 µL) in hybridization buffer [5x
standard saline citrate (0.75 mol/L NaCl, 0.075 mol/L sodium
citrate) containing 1 g/L sodium dodecyl sulfate and 5 g/L
casein] was incubated on the surface for 10 min. The surface was
rinsed with a stream of distilled H2O.
mecA biotinylated detection probe (30 µL of a 1 µmol/L
solution) in hybridization buffer was then added to the surface. After
5 min, the surface was rinsed with distilled H2O,
and 0.9 mg/L anti-biotin antibody labeled with HRP (Jackson
Immunoresearch) was added to the surface for 5 min. After the surface
was rinsed, 35 µL of precipitating tetramethylbenzidine (TMB)
substrate (Kirkegaard and Perry) was added and incubated for 5 min.
After a final rinse, the assay was complete. All incubations occurred
at room temperature (19–25 °C). Total incubation time was 25 min.
capture probe quantification
The mecA capture probe was 5' labeled with
-32P-labeled ATP (ICN) using T4 kinase
(Promega) by standard methods (9). The thin film assay was
carried out, and the counts remaining on the surface were quantified by
Cerenkov counting after each wash step in a Wallac scintillation
counter.
single nucleotide mismatch discrimination
Six mecA target molecules (T1–T6) were synthesized
with single-base mismatches in either the capture or detector
hybridization region (Table 1
). The thin film assay was performed at room temperature.
However, the wash solution was heated to 37 °C. After each assay
incubation step, the 37 °C wash solution was used to rinse the
surfaces. The relative signal intensity of each mismatch target
was compared to control assays with 100% homologous target or no added
target and quantified with the FPE.
array hybridization
The array experiment was performed by hybridization of a
biotinylated oligonucleotide to its surface-bound complement. All steps
in the array hybridization experiment were performed at room
temperature on a 14-mm square thin film surface. The capture
oligonucleotides A, B, or C (100 µmol/L) were adsorbed to the surface
in a 2-µL volume, with 2-mm spacing between each spot. The entire
surface was covered with the biotinylated probes A, C, or A plus C (100
µL of 1 µmol/L) and incubated for 10 min. After the surface was
washed, the anti-biotin HRP-antibody conjugate was added for 10 min,
followed by washing and incubation with TMB substrate for 3 min. Total
assay time was 23 min.
detection of target from serum and urine
The mecA target (1 nmol/L) was equilibrated in human
serum (Sigma), human urine, or water for 1 h at room temperature.
These samples were then diluted 1:1 in 2x hybridization buffer, and
the standard assay protocol was followed. The results were evaluated
both visually and by ellipsometry.
assays incorporating competitor dna
Sheared salmon sperm DNA (10 and 100 mg/L; 5 Prime-3 Prime) was
added to 1 nmol/L target DNA, and the standard assay protocol was
followed.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A). The sample is applied to the
center of the surface, and target detection is enhanced through
deposition of an HRP-catalyzed organic thin film (Fig. 1B
). The
additional thickness alters the antireflective properties of the
surface such that the red and green components of visible light
(500–700 nm) are attenuated, whereas wavelengths corresponding to
purple and blue (400–500 nm) are reflected, producing a color change
from gold to purple.
|
surface construction
The optical surfaces are prepared using a dual ion beam deposition
system to deposit three discrete layers. To produce the proper
intensity of reflected light, antireflective surfaces are designed such
that the refractive index of each layer is approximately the square
root of the one below. Thus, the choice of materials is essential for
proper surface response. In our biosensor, the first layer is amorphous
silicon of 2000 Å with a refractive index of 4.0. Above the silicon
layer, silicon nitride is deposited. Both silicon and silicon nitride
are deposited using ion source bombardment of a pure single crystal
silicon target in an argon or nitrogen atmosphere. Silicon nitride is
chemically inert and has an index of refraction of 2.0 and a nominal
thickness of 150 Å. In the same apparatus, a DLC layer of 200–250 Å,
with a refractive index of 1.9–2.1, is deposited using the
second ion source, which provides radio frequency-assisted ion beam
decomposition of methane gas. The final DLC coating functions both as a
component of the antireflective layer and to mediate capture probe
attachment. The DLC layer is hydrophobic and serves as a suitable
substrate for passive adsorption of nucleic acid or protein capture
molecules. The final hybridization dependent thin film has a refractive
index of 1.4–1.45.
The DLC layer was analyzed for chemical composition by three methods. Auger electron spectroscopy uses a focused electron beam to create secondary electrons near the surface of the sample that have energies characteristic of the element from which they are released. All elements can be detected except hydrogen and helium. We found that the DLC layer was composed of pure carbon, free of contaminants. To determine how the carbon was bound to hydrogen, the DLC was analyzed by secondary ion mass spectroscopy analysis, which bombards the surface with a narrow beam of ions. Ejected ionized species provide a mass spectrum composed of fragment ions of the various functional groups of the surface. The DLC surface composition was mainly ionic fragments of C+, CH+, CH2+, and CH3+.
Further characterization of the DLC surface was provided by XPS. In XPS, monoenergetic soft x-rays bombard the DLC surface and eject electrons, and the presence of the elements on the surface can be directly distinguished by analysis of the kinetic energy of the ejected photoelectrons. XPS revealed the presence of carbon in the DLC coating and confirmed that no contaminants other than a small amount of oxygen were incorporated in the outermost layer of the surface. XPS also allowed examination of the hybridization state of carbon-to-carbon bonds. The ion beam-deposited DLC in this study was composed of ~80% sp3 plus 20% sp2 carbon, indicating that the carbon structure was diamond-like rather than graphite-like.
capture probe characterization
The target-specific surfaces were prepared by applying the
20-nucleotide (nt) capture probe solution to the optically coated
membranes and allowing the spot to dry. No further manipulation was
required to permanently adhere capture probe to the DLC membrane.
Various modifications of the 3' end of the capture probe were tested
for optimal attachment to the optical surface, including fluorescein,
C3, C3 amine, and
C18. These modifications either enhance
hydrophobicity or provide a spacer between the surface and the capture
probe sequence to alleviate potential steric hindrance to
hybridization. The optimal capture probe was a 3'-modified
C3 amine. The amount of capture probe retained on
the surface was quantified by radiolabeling the 5' end with
32P and measuring the counts retained on the
surface at each step in the assay. The application of 10 pmol of
capture probe to the membrane in a 4-mm diameter spot produced
retention of 38 fmol, or 0.4% of the starting material. Therefore, the
density of capture probe was 1825 molecules/µm2
(3 zmol/µm2).
limit of detection
The experimental hybridization assay target was a 60-nt synthetic
sequence from the mecA gene of Staphylococcus
aureus encoding methicillin resistance. The capture probe
contained a 3'-terminal amine, a 3-carbon spacer, 5 nonhomologous nt,
and a 15-nt sequence homologous to the target. The detector probe
contained 15 nt of homologous sequence and 5 nonhomologous nt
terminated with a 5' biotin. The nonhomologous nucleotides served as a
spacer between either the surface and the hybridizing sequences or the
hybrid and biotin. The calculated melting temperature for target
interacting with immobilized capture or detection probe under our
standard assay conditions was 44 and 43 °C, respectively
(9).
Target concentrations of 0.5–1000 pmol/L were tested in the thin film biosensor by incubating the target oligonucleotide for 10 min on the capture probe-coated surface. After hybridization of biotinylated detection probe, the organic thin film was enhanced by binding the HRP-labeled anti-biotin antibody and subsequent precipitation of TMB substrate. Total assay time was 25 min, and the visual limit of detection was 5 pmol/L or 150 amol/sample.
single nucleotide mismatch discrimination
To assess the ability of the silicon biosensor to discriminate
single-nucleotide polymorphisms, six target molecules were synthesized
with single transversion point mutations (Table 1
). The mismatch was
either in the center or near the terminus of the probe hybridization
region. Hybridization efficiency of these targets was compared to that
of the 100% complementary target. All hybridization steps were carried
out at room temperature. Stringency was mediated by rinsing the surface
with 37 °C wash solution after each hybridization step. Assay
results were quantified with the FPE (Fig. 2
). Both the visual appearance of the membranes and the
ellipsometric values demonstrated that this simple selection procedure
allowed discrimination of single-base mismatches between the target
sequence and the capture or detection probe. The intensity values from
the internally located mismatches (T1, T2, and T5) were
indistinguishable from the negative control. The intensity values from
the mismatches located one nucleotide from the end of the hybrid (T3,
T4, and T6) were higher but still below that of the perfectly matched
hybrid. There was no difference in hybridization efficiency if the
mismatch was located in the capture or detection sequence.
|
array hybridization
To investigate the adaptability of the thin film biosensor to
multitarget analysis, we constructed a 3 x 3 array of capture
probes. Capture probes were either 100% complementary (probe A), 80%
complementary (probe B, 12 of 15 nt), or noncomplementary (probe C,
100% mismatch) to a biotinylated target sequence (Fig. 3
). Target detection was enhanced by HRP-catalyzed thin film
formation. All hybridizations and washes were performed at room
temperature. The arrays were probed with multiple target combinations.
In each case, the expected hybridization pattern was generated. There
was no hybridization to capture probe B, the 80% match. The results
were visually interpreted, and only hybridization to perfect
complements was seen.
|
target detection from serum and urine
To test the ability of the thin film biosensor to detect target
nucleic acid from simulated clinical samples, the mecA
target was added to control solution, or human serum or urine. The
samples were stored at room temperature for 1 h. The thin film
assay was then performed and interpreted visually. Target was detected
in both serum- and urine-treated samples with signal strength equal to
the control. No signal interference or false-positive results were
observed. Quantification by ellipsometry demonstrated <5% signal
intensity difference in the results from serum or urine compared with
those of the control.
The effect of competitor DNA on the thin film assay was analyzed by addition of a 5000-fold excess of salmon sperm DNA to the target. No signal difference was observed compared with samples without competitor DNA.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The relationship between color change and molecular thickness can be simulated theoretically and used to predict nonreacted surface color and the increased thicknesses required to produce various colors. For example, with the present layer composition and thickness, we would expect the addition of 10 Å of uniform thin film to alter the surface color from the initial gold to a light purple color. The color change is a function of the alteration in the destructive interference pattern from antireflective to blue (producing a gold appearance) to a higher percentage of blue reflection as the thickness increases. Alternative materials may be modeled and constructed to customize the surface response to a desired thickness and refractive index change.
A key feature of the silicon surface is that it is relatively two-dimensional compared with typical surfaces used for hybridization assays, such as nitrocellulose or polystyrene. The flatness of the biosensor surface may improve on-rates of hybridization and decrease nonspecific binding. Low nonspecific binding creates an assay format with a high signal-to-noise ratio, thus allowing for detection of very low concentrations of analyte (10). The thin film assay is compatible with typical clinical matrices, and assay performance is not affected by competitor DNA. To our knowledge, detection of target from a true clinical matrix has not been reported for other biosensor technologies (3)(4)(5)(6). Biosensor methods based on refractive index changes, such as evanescent field or reflectometric interrogation, cannot discriminate nonspecific adsorption from specific effects (11) and thus may have difficulty with direct examination of clinical specimens.
At present, detection of nucleic acid targets is primarily based on fluorescent or chemiluminescent methods (12). Although these methods are sensitive, they require sophisticated instruments and specialized reagents and are not suitable for field use or for poorly equipped laboratories. The silicon-based biosensors described here can be manufactured at low cost, and the reagents used to generate the thin films are commonly available and relatively inexpensive. Thin films can be generated in a variety of ways, using substrates that produce insoluble enzymatic products or with film-forming particles.
Although the assay described here is qualitative, the thickness of the film generated by the hybridization event can be accurately quantified by ellipsometry. Ellipsometric quantification allowed the discrimination of targets that differed by a single nucleotide. The FPE is capable of detecting changes in thickness of <10Å, yet the instrument is compact and inexpensive (7)(8). Thus, the thin film biosensor can be easily interpreted by either visual or instrumented detection.
Other direct visual detection methods for nucleic acid targets have reported detection limits of 0.1–10 nmol/L (13)(14). The thin film biosensor described here achieved a visual detection limit of 5 pmol/L (150 amol/sample) in under 30 min without target amplification. The format is simple, and the material cost is consistent with current commercial diagnostic formats. The assay lends itself to automation for increased sample throughput and multitarget analysis for such applications as infectious disease panels, antibiotic resistance panels, or allelic discrimination of specific genetic markers.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  H. Li and L. Rothberg Colorimetric detection of DNA sequences based on electrostatic interactions with unmodified gold nanoparticles PNAS, September 28, 2004; 101(39): 14036 - 14039. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Fixe, V. Chu, D. M. F. Prazeres, and J. P. Conde An on-chip thin film photodetector for the quantification of DNA probes and targets in microarrays Nucleic Acids Res., May 17, 2004; 32(9): e70 - e70. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Jenison, H. La, A. Haeberli, R. Ostroff, and B. Polisky Silicon-based Biosensors for Rapid Detection of Protein or Nucleic Acid Targets Clin. Chem., October 1, 2001; 47(10): 1894 - 1900. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Jenison, A. Haeberli, S. Yang, B. Polisky, and R. Ostroff Thin Film Biosensor for Rapid Detection of mecA from Methicillin-resistant Staphylococcus aureus Clin. Chem., September 1, 2000; 46(9): 1501 - 1504. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
