|
|
||||||||
Oak Ridge Conference |
1 Center for Fluorescence Spectroscopy, Department of Biochemistry and Molecular Biology, University of Maryland School of Medicine, Baltimore, MD.
aAddress correspondence to this author at: Center for Fluorescence Spectroscopy, Department of Biochemistry and Molecular Biology, University of Maryland School of Medicine, 725 West Lombard St., Baltimore, MD 21201. Fax 410-706-8408; e-mail lakowicz{at}cfs.umbi.umd.edu.
Abstract
At the Center for Fluorescence Spectroscopy, we have taken advantage of the favorable properties of surface plasmon-coupled emission (SPCE) to improve fluorescence-based immunoassays. SPCE occurs when excited fluorophores near conducting metallic structures efficiently couple to surface plasmons. These surface plasmons, appearing as free electron oscillations in the metallic layer, produce electromagnetic radiation that preserves the spectral properties of fluorophores but is highly polarized and directional. SPCE immunoassays provide several advantages over other fluorescence-based methods. This review explains new approaches to fluorescence immunoassays, including our own use of SPCE for simultaneous detection of more than one fluorescent marker and performance of immunoassays in the presence of an optically dense medium, such as whole blood.
Biological markers in physiologic fluids are very useful tools in clinical diagnostics. Sensitive and reliable detection of specific biomarkers is crucial in disease identification, therapy, and patient screening. Reliable and quick detection of low concentrations of markers is particularly important in a disease such as acute myocardial infarction (AMI). 1 Myoglobin (Myo), although not completely cardiac specific, is one of the earliest markers to increase after AMI (1)(2)(3)(4)(5)(6) and has been recommended for use in combination with other markers, such as creatine kinase-MB, troponin I, and troponin T, as an early diagnostic indicator for AMI (1)(7)(8)(9)(10). The immunoassay technique is widely used in this procedure (11).
Immunoassay based on fluorescence detection is one approach to high-sensitivity detection of biomarkers (12)(13)(14)(15)(16). Different fluorescence detection approaches include polarization (17)(18)(19)(20)(21), resonance energy transfer(22)(23)(24), and time-resolved gated assays based on long-lived lanthanide emission (25)(26)(27)(28). New approaches to fluorescence immunoassays are being developed, including multiphoton excitation (29)(30)(31), with the emphasis on high-throughput immunoassays (32)(33)(34)(35).
The sensitivity of fluoroimmunoassays is typically limited by background fluorescence, which is present in most biological samples and in the optical elements of the instrumentation. In this report, we describe a new immunoassay format that provides increased sensitivity and background rejection by efficient light collection of emissions occurring near the bioaffinity surface. In this approach, a fluorescently labeled reporting molecule (antibody) is bound near a metallic surface, and the binding produces a highly directional and polarized emission. This effect is based on the resonant coupling of excited fluorophores with collective electron motions/oscillations at the interface between a thin metal film (typically silver or gold) and a dielectric bulk material. The so-called surface plasmons comprise an electromagnetic wave confined to the interface between the metal film and the transparent medium, and the electromagnetic wave of the surface plasmons is coupled to oscillations of free electrons in the metal.
A strong evanescent field induced by surface plasmons can excite a layer of fluorophores that extends up to
200 nm above a thin metallic film into the liquid sample. We demonstrated recently that the reverse process is also possible; i.e., excited fluorophores near the metallic layer may induce surface plasmons in the metallic film, which radiate into the glass substrate (36)(37). This radiation occurs at a sharply defined angle and is almost completely polarized. This phenomenon, which we call surface plasmon-coupled emission (SPCE), appears to be closely related to surface plasmon resonance (36)(38)(39)(40). The interaction between the excited fluorophores and the metal causes coupling of the energy quanta to a 2-dimensional electromagnetic wave at the interface between metal and glass. In turn, this electromagnetic wave is coupled to oscillations of electrons in the metal, which can then radiate an electromagnetic field into the glass. The interaction of excited dipoles with metal is a near-field resonance effect occurring without the emission of photons, similar to the Förster resonance energy transfer mechanism; however, detailed studies of the mechanism are still needed.
Surface plasmon resonance occurs when light impinges on a thin metal film through a medium with a higher refractive index. Resonance occurs only at a specific angle of incidence (
SP) that satisfies exactly the wave vectormatching conditions. At this angle of incidence, the surface plasmons are excited, producing a sharp reflectivity decrease (Fig. 1
, right). We found that plasmons created by nearby excited fluorophores may cause radiation into the glass substrate at the surface plasmon angle (
F) that satisfies the wave vectormatching conditions for the emission wavelengths.
|
There are 2 possible experimental configurations for SPCE excitation. First, the sample can be illuminated from the sample side, which is the so-called reverse Kretschmann (RK) configuration. In this case, the excitation cannot generate surface plasmons in the metal surface. The sample can also be illuminated through a prism at the plasmon resonance angle (
SP), which is called the Kretschmann (KR) configuration (Fig. 1
, left). If the incident angle
I =
SP, then the excited surface plasmons induce an evanescent field above the metal film surface. This evanescent field, which is strongly enhanced (up to 80-fold compared with the incident field) by the resonance interaction (41), extends up to
200 nm into the liquid sample. Hence, the KR illumination leads to a strong selective excitation in close proximity to the metal surface. The enhanced excitation field is confined to the evanescent layer, effectively reducing the background from the sample volume matrix. The exceptional sensitivity of SPCE-based assays has been observed in our and other laboratories (42)(43)(44)(45)(46)(47)(48).
Model SPCE Immunoassay
We used SPCE to develop a model affinity assay with labeled goat anti-rabbit IgG antibodies against rabbit IgG bound to a 50-nm thick silver film (48). Rhodamine Red-Xlabeled IgG was placed near the silver surface by its binding to the surface-bound antigen (Fig. 2
, left). The sample was illuminated in the RK configuration, which does not create surface plasmons in response to the incident light. We measured the emission intensity for all accessible angles from the normal axis. The intensity observed through the prism was sharply directed near ±75 degrees (Fig. 2
, right). This value is consistent with that calculated from minimum reflectance for p-polarized plasmon mode for a 595 nm emission wavelength (49). The emission spectrum of the SPCE was characteristic of the rhodamine probe (Fig. 3
, bottom), and the spectrum was not corrupted by scattered light at the excitation wavelength.
|
|
We tested the use of SPCE to monitor the binding kinetics of rhodamine-labeled antibodies to the surface-bound antigen. The emission intensities after addition of labeled antibody are shown in the top panel of Fig. 3
. The emission climbs rapidly and reaches a limiting value. It is important to recognize that this 10-fold change in intensity is not produced by a change in the rhodamine quantum yield on binding but by transport of the rhodamine molecules to be near the metal film. We measured the effect of binding of rhodamine-labeled goat antibody to the antigen while both were free in solution and found that the intensity decrease attributable to binding was
25%. This result further demonstrates that the intensity change is a product of probe localization in the evanescent field near the silver. The use of SPCE thus is a generic method for detection of surface localization by a change in intensity but does not require a change in the fluorophore quantum yield on binding.
Advantages of SPCE-Based Immunoassays
background suppression
We tested several optical configurations to determine the relative intensities and extent of background rejection possible with SPCE (48). These 3 configurations are shown in the right-hand panels in Fig. 4
. The sample consisted of the surface saturated with Rhodamine Red-Xlabeled antibody. AlexaFluor-647labeled antibody solution (which had no affinity to the surface) was added to mimic autofluorescence from the sample. The 0.03 µmol/L concentration of this antibody (0.13 µmol/L Alexa dye) produced dominant free-space fluorescence signal from the sample at 670 nm. The sample was first excited using the RK configuration, and the free-space emission was observed from the same water side of the sample (Fig. 4
, top panel). Compared with subsequent measurements (below), the intensity of the desired rhodamine antibody was weak. The free-space emission was dominated by the emission from Alexa at 670 nm, with only weak rhodamine emission at 595 nm. We then measured the emission spectrum of the SPCE signal (Fig. 4
, middle panel), while still using RK illumination. The emission spectrum was dramatically changed from a 10 to 1 excess of the unwanted background to a 5 to 1 excess of the desired signal. The use of SPCE therefore allowed selective detection of the rhodamine-labeled antibody near the silver film.
|
We next changed the mode of excitation to the KR configuration (Fig. 4
, bottom panel). In this case, the sample was illuminated at
SP, creating an evanescent field in the sample. This increased overall intensity 10-fold and further suppressed the unwanted emission from AlexaFluor-647. The increased intensity and decreased background are the products of localized excitation by the resonance-enhanced field near the metal. In this case, the emission was produced almost entirely by the rhodamine, with just a minor contribution from the Alexa-labeled protein. In our opinion, SPCE-based immunoassays provide unprecedented background rejection with a simple optical configuration and without electronic gating.
multicolor immunoassays
The SPCE phenomenon offers intrinsic wavelength resolution, which can be used in multicolor immunoassays. The experimental configuration for 2-wavelength SPCE (50) is shown in the top panel of Fig. 5
. The silver surface, coated with a mixture of Rhodamine Red-Xlabeled and AlexaFluor-647labeled antibodies, was illuminated at the surface plasmon angle through the glass prism (KR configuration) by 532 nm light, which excites both dyes. The free-space emission was normal to the sample surface, on the side distal from the prism, with a filter and a fiber optic bundle. SPCE was observed on the prism side of the sample at 2 different angles through appropriate long-pass filters for each labeled antibody (595 nm for the Rhodamine Red-Xlabeled antibody and 665 nm for the AlexaFluor-647labeled antibody).
|
To preadjust the optical setup, we calculated the reflectivity curves with available software (51)(52). The reflectivity minima were at 72.5 degrees for 532 nm, and at 69 and 67 degrees for 595 and 665 nm, respectively (Fig. 5
, bottom). We thus expected to obtain excitation of surface plasmons with a 532 nm incident angle of
72.5 degrees and to observe the emissions for the Rhodamine Red-X and AlexaFluor-647labeled antibodies at 69 and 67 degrees, respectively.
We examined the angle-dependent emission intensity for an antigen (rabbit IgG)-covered surface, which was saturated with a mixture of Rhodamine Red-X and AlexaFluor-647labeled antibodies (Fig. 6
, left). The emission from both labeled antibodies was strongly directional at different angles on the prism, as expected. The emission from rhodamine peaked at 71 degrees and that from Alexa dye at 68 degrees.
|
We used SPCE at 2 observed angles to simultaneously measure the binding kinetics of both labeled antibodies (Fig. 6
, right). The binding kinetics were similar, although the final intensities were different (Fig. 6
,
and
). The binding was also measured by use of KR excitation and the free-space emission (Fig. 6
, and
). The intensities were >10-fold higher for SPCE than for the free-space emission. The SPCE-based immunoassay therefore provides better resolution for a multicolor/multianalyte system than a system for which free-space emission is observed. The SPCE immunoassays offer the opportunity to simultaneously observe of multiple fluorescent markers.
spce immunoassay in optically dense media
In medical testing, it is often desirable to perform homogeneous assays without separation steps, sometimes in whole blood. We reasoned that an SPCE signal should be detectable in optically dense media because this signal arises from the sample within 200 nm of the surface.
In the endpoint experiment, AlexaFluor-647labeled anti-rabbit IgGs were first bound to rabbit IgGs immobilized near the silver surface (Fig. 7
, left) (53). The excess of nonbound antibodies was then washed away, and a sample matrix was added (serum, whole blood, or blocking solution for comparison), followed by measurements of the fluorescent signal and spectrum. We found that the emission was strongly directional and focused near an angle of 58 degrees. The spectrum of the SPCE (in KR configuration) was characteristic of the AlexaFluor-647 probe and not corrupted by scattered light at the excitation wavelength for all tested sample matrixes (Fig. 7
, right). The whole blood sample used for this measurement had an absorbance of
2.5 at 665 nm (emission maximum of the tested dye AlexaFluor-647) at the optical pathlength of 0.2 mm (as used for the SPCE experiment). This absorbance in a 0.2-mmthick blood sample would attenuate the fluorescence signal
102.5-fold under free-space conditions. As shown in the right-hand panel of Fig. 7
, with SPCE generation and detection, the signal was attenuated <2-fold in a human serum sample and
3-fold in a whole blood sample compared with the transparent buffer medium (blocking solution). These results demonstrate the potential of SPCE in optically dense samples.
|
To characterize the effect of serum and whole blood on the SPCE signal, we estimated the limit of detection (LOD) as the concentration of labeled IgG (binding to the immobilized antigen) giving the same SPCE signal as the blank measurement (0 concentration of labeled antibodies) plus 3 SD, assuming linearity between 0 and 70 nmol/L (10 mg/L) of labeled IgG. In our model immunoassay, we estimated a 3- to 5-fold increase in LOD for serum and a 5- to 10-fold increase in LOD for whole blood.
In addition, we were able to monitor the binding process, which itself could be affected by the presence of serum or whole blood (Fig. 8
). The increase of the signal evidently can be correlated with specific binding, as shown in Fig. 8
. Nonspecific binding of the labeled anti-rabbit IgG to the wrong antigen (goat IgG) leads to an SPCE signal that does not change with time. We believe that the ability to perform assays in whole blood can speed up and simplify clinical assays.
|
SPCE Myo Immunoassay
To demonstrate usefulness of the SPCE-based immunoassay format for detection of a cardiac marker (54), we performed the assay schematically depicted in Fig. 9
(left), in which Rhodamine Red-Xlabeled antibody was bound to Myo immobilized near the silver surface. The emission spectrum of the SPCE was characteristic of the Rhodamine Red-X probe (Fig. 9
, right). A remarkable characteristic of SPCE is almost complete polarization in the p direction, meaning that the electric vector is oriented parallel to the plane of incidence. The right-hand panel of Fig. 9
shows the emission spectra collected through an emission polarizer (oriented p or s). This p polarization demonstrates that the emission is the product of surface plasmons that under these conditions cannot emit s-polarized light. An emission polarizer in the p orientation can be used to further suppress part of the free-space emission that can be transmitted through the metal film, some of which will display s polarization. We measured the sample emission with RK and KR configurations. The spectrum of Rhodamine Red-Xlabeled antibody was the same in both cases.
|
For the KR configuration, the 532 nm excitation was at an angle of 74.5 degrees, which we found yielded the highest SPCE fluorescence. The emission was strongly directional at the angle 72 degrees. With these optical conditions, we measured the binding kinetics of the Rhodamine Red-Xlabeled anti-Myo antibodies to the surface-bound Myo at various Myo concentrations. The SPCE emission intensities after addition of labeled antibody are shown in Fig. 10
. The signal rapidly increases during the first few minutes and grows slowly during the next 10 min. As shown in Fig. 10A
, the background signal (0 µg/L), representing nonspecific binding, is definitely much lower than for the lowest Myo concentration used, 50 µg/L, which is below the clinical cutoff Myo concentration of 90 µg/L for healthy patients.
|
We also examined the influence of human serum and 17% hemoglobin solution on the sensitivity of our Myo assay in a 1.0-mmthick sample (not shown). Because signal attenuation was only 2- and 3-fold in serum and hemoglobin solution, respectively, a subclinical Myo concentration at 50 µg/L was readily detectable in those optically dense media.
Analytical assay characteristics, such as precision, accuracy, specificity, LODs, limits of quantification, and linearity, may depend on numerous factors, including analyteantibody characteristics (such as Myo concentration and antibody affinity), assay conditions (such as temperature and incubation time), and such characteristics as slide substrate (metal layer coverage and protective SiO2 layer) property variations. We demonstrated the applicability of SPCE technology to a real immunoassay within a clinically important cardiac marker concentration range. We did not seek to optimize the assay, which will be the topic of further investigation; however, the accuracy of the SPCE assay related solely to signal detection can be estimated. The variation in measured Myo concentration because of the signal detection accuracy, which was 2% or better, was <5% at Myo concentrations of 50 and 125 µg/L and <10% at Myo concentrations of 500 and 1000 µg/L, based on the semilog trend line covering the full range of Myo concentration studied (Fig. 10C
).
Conclusions
SPCE immunoassays using the novel plasmonic approach described here may provide the following advantages over other fluorescence-based methods:
The proposed approach requires no washing steps and simplifies complex, heterogeneous assays. The SPCE measurements do not involve any sophisticated laser light sources and can be done with arc lamps or light emission diodes (LEDs). SPCE is also efficient because light excitation is not needed for the coupling of excitation to the surface plasmons; electroluminescence can lead to an efficient SPCE (55).
Acknowledgments
This work was supported by the National Center for Research Resources (Grant RR-08119) and the National Institute of Biomedical Imaging and Bioengineering (Grant EB-00682).
Footnotes
1 Nonstandard abbreviations: AMI, acute myocardial infarction; Myo, myoglobin; SPCE, surface plasmon-coupled emission; RK, reverse Kretschmann; KR, Kretschmann; and LOD, limit of detection. ![]()
References
The following articles in journals at HighWire Press have cited this article:
![]() |
Y. Luo, M. Chen, Q. Wen, M. Zhao, B. Zhang, X. Li, F. Wang, Q. Huang, C. Yao, T. Jiang, et al. Rapid and Simultaneous Quantification of 4 Urinary Proteins by Piezoelectric Quartz Crystal Microbalance Immunosensor Array Clin. Chem., December 1, 2006; 52(12): 2273 - 2280. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |