Clinical Chemistry 43: 379-383, 1997;
(Clinical Chemistry. 1999;43:379-383.)
© 1999 American Association for Clinical Chemistry, Inc.
Optical tweezers-based immunosensor detects femtomolar concentrations of antigens
Kristian Helmerson1,a,
Rani Kishore1,
William D. Phillips1 and
Howard H. Weetall2
1
Atomic Physics Division, and
2
Biotechnology Division, National Institute of Standards and Technology, Gaithersburg, MD 20899.
a Author for correspondence. Fax 301-975-3038; e-mail kristian{at}enh.nist.gov
 |
Abstract
|
|---|
We used optical tweezers (optical trapping technology) to measure the
force required to separate antigen–antibody bonds. Under
competitive-binding conditions, we used the force determination to
detect and measure protein antigen concentrations as small as 1 fmol/L
(10-15 mol/L).
Key Words: indexing terms: lasers • antigen–antibody binding • immunoassays
|
Introduction
|
|---|
Detection and quantification of ultralow concentrations of
analytes are becoming increasingly important (1). Methods
that do not require amplification techniques to detect infectious
organisms, viruses, and nucleic acid targets could find wide
application in the clinical environment. To this end, we have carried
out a series of experiments leading to the detection of femtomolar
concentrations of a protein antigen by a method we believe could be
widely applicable and capable of automation. This approach involves
optical trapping technology capable of sensing single antigen–antibody
bonds. Using a competitive-binding or displacement-type assay, we can
detect extremely small quantities of a soluble antigen added to the
system.
|
Materials and Methods
|
|---|
principle
We have constructed a sensor that is based on optical trapping
technology, i.e., optical tweezers. Optical tweezers are focused laser
beams used to trap and remotely manipulate dielectric particles,
including cells and other biological objects (2)(3)(4). The
change in momentum of the light transmitted by the dielectric object
results in a force that traps objects having an index of refraction
greater than that of the surrounding medium at the local maximum of the
intensity of the electromagnetic field, i.e., at the focus of the laser
beam. Svoboda and Block (4) have reviewed the
principles of optical forces as well as the various configurations and
applications of optical tweezers.
Figure 1
illustrates the basic principle of our device. We use optical
tweezers to trap a microsphere coated with an antigen and then pull the
microsphere away from a surface coated with the corresponding antibody.
Throughout, we measure the force applied by the optical tweezers to
break the antigen–antibody bonds and to pull the microsphere away from
the surface. To detect the presence of small quantities of the antigen
in solution, we use a competitive-binding displacement approach: The
binding of the free antigens in solution to the antibodies on the
surface is detected as a decrease in the average force required to pull
the antigen-coated microsphere away from the surface.
View larger version (71K):
[in this window]
[in a new window]
|
Figure 1. Basic scheme of our optical tweezer-based immunosensor.
An antigen-coated microsphere is trapped and pulled away from an
antibody-coated surface by use of the optical tweezers. The minimum
amount of force applied by the tweezers to break the
microsphere-coupled antigen–antibody bonds is measured. Detection of
free antigens in solution is manifested as a reduction of this applied
force, the result of displacement of microsphere-coupled antigen by
free antigen in the binding to the antibody.
|
|
apparatus1
Our optical tweezers apparatus consists of a Zeiss inverted
microscope equipped with a 100x (numerical aperture 1.4) oil-immersion
objective lens, a video camera and monitor, a computer-driven
translation stage capable of motion in three orthogonal directions, and
a continuous-wave Nd:YAG laser emitting light at a wavelength of 1.06
µm. The laser light is coupled into the back of the objective lens
with a dichroic mirror, which enables us to simultaneously view and
trap the microspheres. Transparent, polarizable objects with an index
of refraction higher than the surrounding medium, such as these
microspheres in buffer, are trapped at the local maximum of the
intensity of the laser light, tightly focused by the objective lens.
Typically, the size of the focal spot is kept fixed; therefore, the
strength of the trapping force is proportional to the power of the
laser.
The microspheres, suspended in a buffer solution, were contained in a
chamber constructed from a glass microscope slide with a 1.0-cm hole
drilled through it and two glass cover slips on each side. The
coverslip on the objective-lens side of the chamber contained the
silane-coupled antibodies. Both cover slips were sealed to the
microscope slide with silicone vacuum grease. The total volume of the
chamber was ~100 µL.
sample preparation
The antigen used in these experiments, bovine serum albumin (BSA,
98-99%; Sigma Chemical Co., St. Louis, MO), was covalently coupled to
4.5-µm-diameter latex microspheres with carboxyl groups (Bangs Labs.,
Carmel, IN). The covalent coupling was accomplished by first removing
0.1 mL of the original 50 g/L suspension of microspheres and washing
with pH 6.6 buffer (0.05 mol/L potassium phosphate, 0.1 mol/L sodium
chloride, 2 g/L gelatin, and 0.1 mL/L thimerosal). Next, we added 1.0
mL of 50 mmol/L 2-(N-morpholino)ethanesulfonic acid
buffer (Sigma Chemical Co.), pH 5.5, to the desired quantity of BSA
plus microspheres and vortex-mixed. A freshly prepared 10 g/L solution
of water-soluble 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (Sigma
Chemical Co.) was added to the microspheres and stirred for 4 h.
Unreacted carbodiimide was removed by centrifuging and washing the
microspheres three times at 15 °C with phosphate-buffered saline
(PBS; 0.05 mol/L potassium phosphate and 0.1 mol/L sodium chloride), pH
7.5, containing 0.1 g/L thimerosal. We stored the washed microspheres
in PBS at 4 °C. For the experiments reported here, we diluted the
BSA-coupled microspheres to ~0.1 µmol/L, which yielded ~1000
microspheres in the sample chamber.
The corresponding antibodies, mouse monoclonal anti-BSA antibodies,
were covalently attached to glass cover slips through silane coupling.
Silane-coated cover slips were prepared by first boiling the cover
slips in 10% nitric acid for 1 h and then washing with distilled
water until the pH of the water was neutral. Silane solution was
prepared by adding 5 mL of 3-glycidoxypropyltrimethoxysilane and 5
mL of tetramethylorthosilicate (both from Aldrich Chemical Co.,
Milwaukee, WI) to 100 mL of deionized water. The pH of the silane
solution was adjusted to 4.0 with 10% acetic acid solution. The cover
slips were dipped in the silane solution, dried at room temperature,
heated for 90 min in an oven at 110 °C, and then mounted on glass
slides. To the coverslip surface we added 100 µL of 0.05 mol/L
potassium phosphate, pH 8.0, followed by 10 µL of 2.8 g/L anti-BSA
monoclonal antibodies (Sigma Chemical Co.). After incubating the slides
at 5 °C for 72 h, we rinsed the coverslips with the
PBS–thimerosal solution.
measurement of binding
We detected the binding of the microsphere to the coverslip
surface as follows. Microspheres resting on the surface of the
coverslip were located optically with the microscope. We focused the
microscope objective on the surface of the coverslip and then
positioned the objective in the center of the microsphere in the plane
of the coverslip. Using the computer-driven stage, we displaced the
objective lens 5.0 µm toward the coverslip surface. This corresponds
to placing the focus of the laser beam about one microsphere radius
above the middle of the microsphere, into the chamber, away from the
coverslip surface. We then slowly increased the laser power from zero
until the point at which the microsphere could be seen to jump away
from the coverslip surface into the focus of the laser beam. The
minimum power at which the microsphere was pulled into the optical trap
was recorded. The laser power was typically increased over a timespan
of 5 s. However, the minimum laser power required to lift the
microsphere off the surface and into the trap was unchanged when we
increased or decreased this interval twofold (data not shown).
|
Results and Discussion
|
|---|
specific vs nonspecific binding experiment
We performed an initial series of experiments to study specific
and nonspecific binding of antigens to the silanized surfaces, i.e.,
with and without antibodies, respectively. Two series of measurements
were made, involving microspheres coated with BSA and silane-coated
cover slips with and without anti-BSA coupled to them. For each series
of measurements, we varied the surface coverage of the microspheres by
changing the amount of BSA offered during the coupling procedure (from
1.45 x 10-7 mol/L to
1.45 x
10-15
mol/L). Fig. 2
presents the titration data for the BSA-coated microspheres,
i.e., the laser power required to pull a microsphere, coated at a
particular concentration of BSA, off the coverslip surface (mean
± SD, n = 10).
View larger version (52K):
[in this window]
[in a new window]
|
Figure 2. Power required to pull microspheres coated with BSA off of
a silane-coated glass coverslip (light bars) and off of a
silane-coupled anti-BSA-coated coverslip (dark bars).
The increase in the measured laser power corresponds to increasing
binding force between the BSA-coated microspheres and the silanized
surface. The increase in the binding force with increasing BSA on the
surface of the microspheres is interpreted as arising from the
increasing number of BSA—anti-BSA bonds and of BSA–methyl-terminated
silane bonds for the anti-BSA-coupled and noncoupled silane-coated
surfaces, respectively. The BSA concentrations indicated are those used
in the coupling reactions and not the quantities bound; therefore, they
represent the maximum quantity of protein that could be coupled to the
microspheres. The mean and SD of 10 measurements made with independent
microspheres are shown.
|
|
As Fig. 2
shows, microspheres treated with BSA concentrations of
1.45 x 10-13 mol/L
(0.145 pmol/L) or less bind to the coverslip surface with a force
essentially the same as when there is no BSA on the microsphere. For
microspheres coupled at 1.45 pmol/L or greater BSA concentrations, the
binding force of the microsphere to the silanized surface increases
with increasing BSA concentrations; however, the measured binding force
of BSA to anti-BSA is much larger than the measured binding force of
BSA to a coverslip coated with silane only. We interpret this as
arising from both the specific and nonspecific binding of the BSA to
the silanized surfaces with and without anti-BSA, respectively.
Given the number of microspheres used in the coupling reaction, and
assuming 100% coupling of available BSA molecules to the microspheres,
we calculate that the average number of BSA molecules coupled to the
microspheres at the BSA concentration of 0.145 pmol/L is 1. (This
estimate of the number of BSA molecules coupled to the microsphere is
probably good within an order of magnitude. In practice, there will be
<100% coupling of the BSA to the microspheres; however, loss of
microspheres during washing will tend to compensate for the reduced
coupling.) Thus, we expect that microspheres treated at BSA
concentrations <0.145 pmol/L should have essentially no BSA available
to bind to the coverslip surface. As the number of BSA molecules on the
microsphere increases, the number of BSA-to-surface bonds for both
specific and nonspecific binding should increase, along with a
corresponding increase in the measured binding force.
The high selectivity provided by the molecular recognition of
antibodies for antigens is shown in the data for Fig. 2
. In each case
in which microsphere-coupled BSA exhibited binding, the binding force
of the BSA to a silane-coupled anti-BSA surface was larger than the
binding force of the BSA to the silane surface without anti-BSA.
Despite the strong tendency of BSA to bind to surfaces terminated with
methyl groups (e.g., our nonfunctionalized silane surface
(5)), the measured binding force of BSA to anti-BSA was
30% to 50% larger for microspheres containing coupled BSA at
concentrations of 1.45 x 10-12 mol/L to 1.45 x
10-7 mol/L, respectively. Thus, Fig. 2
suggests that the
increase in the specific binding of a microsphere coupled with BSA to
the anti-BSA-coated silane surface occurs for microspheres coupled at
BSA concentrations between 1.45 x 10-13 and
1.45 x 10-12 mol/L. In this range, the measured
increase in the binding force would arise from a single
antigen–antibody binding pair or at most a few such pairs.
In an additional experiment we tested our conclusion that the increased
force observed with the BSA-coupled microspheres on the anti-BSA-coated
coverslip was attributable to specific binding and not to the presence
of any nonspecific IgG–BSA interaction. Because mouse IgG recognizes
mouse serum albumin and does not cross-react with BSA, we coupled
nonspecific mouse IgG (Sigma Chemical Co.) to a silanized coverslip and
repeated the binding force measurement experiments with microspheres
prepared at BSA concentrations ranging from 1.45 x
10-7 to 1.45 x 10-15 mol/L. The results
showed that the force required to break the microsphere-to-surface
bonds remained constant across the entire range of BSA concentrations
coupled to the microspheres. This is consistent with the interpretation
that the increased binding force observed in the presence of the
specific antibodies results from the specific antigen–antibody
interaction. In contrast with the data of Fig. 2
, the absence of any
increase in the nonspecific binding force of the BSA-coated microsphere
with the silanized coverslip as the coverage of the microsphere
increased suggests that the coverslip was, effectively, fully coated
with the nonspecific mouse IgG. This would correspond to an average
spacing between nonspecific IgG molecules of, at most, the radius of
the microsphere. Hence, the minimum average surface density of
nonspecific IgG is 2 x 107
molecules/cm2.
immunosensor experiment
To demonstrate use of the device as an immunosensor, we performed
a typical competitive or displacement-type assay, using a BSA-coated
microsphere (BSA concentration 1.45 x 10-7 mol/L)
and a silane-coupled anti-BSA-coated surface. For each experiment, we
added 1 µL of buffer solution (containing free BSA at various
concentrations) to the chamber containing the microspheres in 100 µL
of buffer and incubated this for 2 h at room temperature. We then
measured the binding force of the microspheres to the anti-BSA-coated
surface as described earlier. Fig. 3
shows the results of our experiments as a function of free BSA
in solution.
View larger version (56K):
[in this window]
[in a new window]
|
Figure 3. Power required to pull microspheres coated with BSA (at
1.45 x 10-7 mol/L) off of a silane-coupled
anti-BSA-coated glass cover slip in the presence of free BSA in
solution.
The increase in the measured laser power corresponds to increasing
binding force of the BSA-coated microspheres with the silanized
surface. The decrease in the binding force with increasing free BSA
concentrations is interpreted as arising from the decreasing number of
anti-BSA binding sites available to the BSA-coated microsphere because
of displacement by the free BSA in solution. The BSA concentrations
indicated are those of free BSA in the chamber. Shown are the mean and
SD of 10 measurements made with independent microspheres.
NaB, measurements made with BSA-coated microspheres and a
surface coated only with silane (no anti-BSA).
|
|
The data in Fig. 3
indicate that the range of sensitivity of our assay
covers at least three orders of magnitude. We observe an increase in
the binding force of the microsphere to the surface for concentrations
of BSA in solution between 1.45 x 10-12 and
1.45 x 10-15 mol/L. The difference in the binding
force of the microsphere to the surface for 0 BSA in solution and for
1.45 x 10-15 mol/L BSA in solution shows that
the assay is sensitive to femtomolar concentrations of antigens. At BSA
concentrations of 1.45 x 10-11 mol/L and higher in
solution, the binding force of the microsphere to the anti-BSA-coated
surface is indistinguishable from the binding force of the microsphere
to a silanized surface without anti-BSA (the NaB value in Fig. 3
). We
interpret this as arising from a complete displacement of the
microsphere-coupled BSA–anti-BSA reaction by the free BSA. An upper
limit for the average surface density of anti-BSA can be estimated,
based on the minimum concentration for complete displacement. Because a
concentration of 1.45 x 10-11 mol/L of BSA in
solution corresponds to ~9 x 108 molecules, the
maximum average surface density of anti-BSA is 9 x
108 molecules/cm2. However, this value is an
overestimation because, in equilibrium, some fraction of the BSA is not
bound but in solution. According to the data in Fig. 2
, the increase in
the difference between specific and nonspecific binding force for BSA
concentrations of 1.45 x 10-12 and 1.45 x
10-7 mol/L is only threefold, even though the BSA surface
coverage of the microsphere has increased by 105. This
implies that the number of BSA to anti-BSA bonds is limited by the
surface coverage of anti-BSA on the silanized coverslip and is at least
threefold higher than the lower estimate based on the results with the
nonspecific mouse IgG. We conservatively estimate the average surface
density of anti-BSA to be between 6 x 107
and 9 x 108 molecules/cm2.
At a concentration of 1.45 x 10-15 mol/L, only
~105 molecules of BSA are in solution. A possible
explanation for the detection of such a low concentration of analyte is
that the dielectrophoretic force (6) of the laser acting
on the free BSA molecules can effectively concentrate the free BSA in
the region of contact between the microsphere of interest and the
surface. In the presence of a gradient electric field, the interaction
of the induced dipole moment of a molecule with the field results in a
dielectrophoretic force on the molecule—essentially the same force
responsible for trapping the microspheres by the optical tweezers. A
simple calculation shows that, in the presence of the dielectrophoretic
force of the laser field used in these experiments, the time required
for a free BSA molecule (with a diffusion constant of 0.59
cm2/s) to diffuse a distance of 100 µm (approximately the
average distance between free BSA molecules at 1.45 x
10-15 mol/L) to the focus of the laser beam is close to
the timescale in which we make our measurements (5 s). The
polarizability of the glass coverslip should also enhance the electric
field at the point through which the laser beam passes and increase the
dielectrophoretic force at the location where the microsphere binds to
the surface, thereby further concentrating the free BSA. Although our
simple estimate of the timescale for bringing the free BSA molecules to
a location where they could bind to anti-BSA and block the microsphere
from sticking is reasonable, a more detailed calculation would include
the binding of the BSA to the anti-BSA. Such a calculation, however, is
beyond the scope of this paper, but further studies of this interesting
dielectrophoretic effect are warranted.
In conclusion, we have demonstrated an optical tweezers-based
immunosensor capable of detecting femtomolar concentrations of antigen
in a competitive-binding assay. Although we have shown that the system
described is capable of such highly sensitive measurements, an
investigation of the limits of sensitivity as well as the development
of some form of sample processing and microfluidics will be necessary
for maximum utilization of this technology in the clinical environment.
We believe the sensitivity of this technology can be improved further
than has been demonstrated here. The broad potential of this approach
for detection and quantification of binding pairs is under further
investigation.
|
Acknowledgments
|
|---|
We thank Joseph Hubbard, William Yap, Barbara Levin, and Baldwin
Robertson for their helpful discussions during preparation of this
manuscript. We also gratefully acknowledge Lori Goldner and Patricia
Purdue for their aid during initial stages of the experiments and
Brooke Bevis for assistance during later stages. Funding was provided
by the National Institute of Standards and Technology (NIST), National
Science Foundation Grant PHY9312572, and the Summer Undergraduate
Research Fellowship Program of the Physics Laboratory of NIST.
|
Footnotes
|
|---|
1 Certain commercial materials and products are identified in this paper to adequately specify the experimental procedures. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology. 
|
References
|
|---|
-
Morgan CL, Newman DJ, Price CP. Immunosensors: technology and opportunities in laboratory medicine [Review]. Clin Chem 1996;42:193-209.
[Abstract/Free Full Text]
-
Ashkin A, Dziedzic JM. Optical trapping and manipulation of viruses. Science 1987;235:1517-1520.
[Abstract/Free Full Text]
-
Ashkin A, Dziedzic JM, Yamane T. Optical trapping and manipulation of single cells using infrared laser beams. Nature 1987;330:769-771.
[Medline]
[Order article via Infotrieve]
-
Svoboda K, Block S. Biological applications of optical forces. Annu Rev Biophys Biomol Struct 1994;23:247-285.
[ISI][Medline]
[Order article via Infotrieve]
-
Silin V, Weetall HH, Van Derah D. SPR studies of the
non-specific adsorption kinetics of human IgG and BSA on gold surfaces
modified by self-assembled monolayers. J Colloid Interface Sci
1997;185: (in press)..
-
Pohl HA. Dielectrophoresis 1978 Cambridge University Press Cambridge. .
Top
Abstract
Introduction
