Silver Particles Enhance Emission of Fluorescent DNA Oligomers

Abstract

Here we describe a new opportunity in methodology for increasing the detectability of fluorescently labeled DNA on solid substrates. We show that the use of glass substrates coated with metallic silver particles results in an approximate 5-fold increase in the intensity of Cy3- or Cy5-labeled DNA oligomers. Proximity to these silver particles also increases the photostability of Cy3- and Cy5-labeled oligomers. These results suggest the use of DNA array substrates with silver particles for increased sensitivity in genetic analysis.

INTRODUCTION

DNA arrays are widely used in studies of gene expression (2,5,16). Detection of hybridized DNA is usually performed using fluorescently labeled DNA. Since the labeled DNA is only present on the surface, the fluorescent signal is limited. Methods that increase the total emission per labeled DNA molecule would provide increased sensitivity, a wider dynamic range, and fewer problems with unwanted background from the reagents, substrates, or optical components.

We suggest the use of metal-enhanced fluorescence to improve detectability on DNA arrays. Metal-enhanced fluorescence is related to surface-enhanced Raman scattering (SERS). In the absence of metal surfaces, Raman scattering is a weak phenomenon and thus difficult to accomplish with dilute samples. However, Raman signals can be dramatically enhanced by the interaction of the target molecules with rough metallic surfaces, typically roughened silver electrodes or irregular silver particles (7,21). The phenomenon of SERS increases the scattering intensity by a factor of 10⁶ for averaged signal and by factors as high as 10¹⁴ for individual molecules on selected metallic particles (8,19). The SERS effect appears to be due to both the electromagnetic interactions of the molecules with the conducting surface and the specific interactions of surface-adsorbed molecules (4,22).

Electromagnetic interactions also occur between fluorophores and conducting metallic surfaces and particles (1,3,10). In the case of fluorescence, the molecules adsorbed directly on the surface are thought to be quenched. Thus, through space, electromagnetic interactions between fluorophores and the metallic surface (metal) are likely to be the origin of fluorescence spectral changes. While theoretical publications often focus on SERS, these same papers often describe the predicted effects on fluorescence. A wide range of effects are expected, including increased and decreased quantum yields, increased and decreased lifetimes, changes in photostability, and increased distances for resonance energy transfer (RET). Given the ubiquitous use of fluorescence in biotechnology and DNA analysis, we have investigated the effects of silver particles on a variety of fluorophores and RET (12–15).

In this report, we focused on whether metallic silver particles can cause useful spectral changes for labeled DNA under the conditions used on DNA arrays. In our experiment, we used dsDNA labeled with either Cy3 or Cy5 and tethered to amino-coated quartz slides, which were half covered with silver island films. Silver island films consist of a layer of sub-wavelength size silver particles that cover about 20% of the surface and do not form a continuous silver coating. These films display the plasmon resonance typical for colloidal metal (9) and have a blue-green color but are not reflective.

MATERIALS AND METHODS

Sample Preparation

The labeled oligomers containing Cy3 or Cy5 on the 5′ ends (Figure 1) were obtained from Synthetic Genetics (San Diego, CA, USA). The complementary unlabeled oligonucleotides were obtained from the Biopolymer Core Facility of the University of Maryland School of Medicine.

Figure 1.

DNA structures and sample geometry. Structures and sequences of the labeled and unlabeled DNA oligomers (top). Absorption spectrum of silver islands on APS and experimental geometry (bottom).

The dsDNA samples (Cy3-DNA or Cy5-DNA) were prepared by mixing the complementary oligonucleotides in 3× SSC buffer to a final concentration of 2 µM. The samples were then heated to 70°C for 2 min, followed by slow cooling. Concentrations were determined using ε(548 nm) = 150 000 M⁻¹ cm⁻¹ for Cy3 and ε(648 nm) = 215 000 M⁻¹ cm⁻¹ for Cy5. The quantum yields were calculated using rhodamine B in water (Q = 0.48) as a reference. The quantum yields of Cy3-DNA and Cy5-DNA in the buffer solution were found to be 0.24 and 0.20, respectively.

We used quartz slides to minimize background emission in our measurements. The entire surface of each slide (1 × 4 cm) was coated with amino groups using 3-aminopropyltriethoxysilane (APS; Sigma, St. Louis, MO, USA). For this purpose, the slides were rigorously cleaned, soaked in a 0.1% aqueous solution of APS for 10 min, and rinsed with water.

Silver island films were formed on half of the amino-coated slides, and the other half was left as an unsilvered control. Silver was deposited by the reduction of silver nitrate using D-glucose, as described previously (13,18). The particles obtained were typically 100–300 nm across, 60 nm high, and covered about 20% of the treated surface. These islands displayed the typical surface plasmon resonance absorption with a peak near 440 nm (Figure 1).

Labeled dsDNA samples (Cy3-DNA or Cy5-DNA) previously hybridized in solution were deposited on the APS-treated slides by placing 250 µL 2 µM solution on a 1 × 4 cm area. The slides were illuminated with a shortwave UV lamp (model UVGL; 25 W for 15 min at a distance of 3 cm) to provide DNA crosslinking to the amino-coated surface and then washed extensively with buffer. The DNA-coated surface was maintained in buffer by using a 0.2-mm demountable cuvette (Starna Cell, Atascadero, CA, USA) that was not coated with amino groups or silver. We attempted the absorption measurements of Cy3-DNA and Cy5-DNA on the unsilvered and silvered areas. We did not notice any significant (more than 20%) difference in dye absorptions, and we believe that the surface concentration of the tethered probes did not differ more than 20% on the silvered and unsilvered areas.

Fluorescence Measurements

For steady-state and time-resolved fluorescence measurements, the sample was placed diagonally with an approximate 45° incident illumination and observation (Figure 1, bottom). Emission spectra were obtained using an SLM 8000 spectrofluorometer and 514 and 605 nm excitation for Cy3 and Cy5, respectively. The emission was selected using 530 and 630 nm long pass filters for Cy3 and Cy5, respectively. Lifetime measurements were performed on a 10-GHz frequency-domain fluorometer (11) with picosecond resolution. For the Cy3-DNA sample measurement, the excitation source was a mode-locked argon ion laser, 514 nm, at about 76 MHz pulse repetition rate. For Cy5-DNA, we used a cavity-dumped R6G dye laser, 605 nm, at 3.8 MHz repetition rate. The same laser light sources were used for the steady-state and frequency-domain measurements and for photographs. The frequency-domain data were obtained with magic angle polarizer conditions. The emission of Cy3 was observed through a combination of a 565-nm interference filter and a 530-nm long pass filter, and Cy5 was observed with a 665-nm interference filter combined with a 630-nm long pass filter.

RESULTS AND DISCUSSION

Figure 2 shows the emission spectra of Cy3-DNA and Cy5-DNA on slides. Depending on the particular silver-coated slide, we observed 5- to 10-fold increases in intensity from the silver-coated part of the slide, as compared with the unsilvered part of the slide. Figure 2 also shows photographs that illustrate Cy3-DNA and Cy5-DNA emission on quartz and silver island films.

Figure 2.

Emission spectra of Cy3-DNA (top) and of Cy5-DNA (bottom) on APS-treated slides, with and without silver island films. The photograph shows the fluorescence spots on quartz and silver taken through 530 and 630 nm long pass filters for Cy3-DNA and Cy5-DNA, respectively.

It is well known that the quantum yield of a fluorophore is given by

$$Q_0=\frac{\mathrm{\Gamma }}{\mathrm{\Gamma }+k_{\mathit{\text{nr}}}}$$ [Eq. 1]

where Γ and k_nr are the radiative and non-radiative decay rates, respectively. The lifetime is given by

$${\tau }_0=\frac{\mathit{1}}{\mathrm{\Gamma }+k_{\mathit{\text{nr}}}}$$ [Eq. 2]

Increases in fluorescence intensity are frequently observed because of the decreases in k_nr, which often occur when fluorophores are immobilized relative to the free solution. Therefore, Figure 2 is a lower value of k_nr on the unsilvered part of the slide. A decrease in k_nr would result in an increase in the lifetime (Equation 2).

We measured the frequency-domain intensity decays of Cy3-DNA and Cy5-DNA on the unsilvered and silvered regions of the slides (Figure 3). The lifetimes are dramatically shortened on the silvered part of the slides, which is better visualized on the time-domain representation of the decay data (Figure 3 and Table 1). An increase in intensity and a decrease in lifetime indicate an increase in the radiative decay rate Γ to a larger value near the silver particles (Γ_m). Increases in radiative decay rates are expected for fluorophores at appropriate distances from silver particles (1,3,10).

Figure 3.

Frequency-domain intensity decays of Cy3-DNA and Cy5-DNA on APS-treated slides, with (●) and without (○) silver island films. The right panels show the time-domain representation of the frequency-domain data.

Table 1.

Decay in Intensity of Emission from Cy3-DNA and Cy5-DNA on APS-Coated Quartz Slides

Sample	τ̄(ns)	<τ>(ns)	αi	fi	τi (ns)	${\chi }_{\mathrm{R}}^2$
Cy3-DNA, Q	1.345a	1.057b	0.362	0.111	0.325
			0.638	0.889	1.472	1.4
Cy3-DNA, S	0.532	0.125	0.674	0.151	0.028
			0.269	0.394	0.183
			0.057	0.455	1.003	1.9
Cy5-DNA, Q	1.560	0.833	0.614	0.159	0.215
			0.386	0.841	1.815	1.4
Cy5-DNA, S	0.411	0.041	0.923	0.438	0.019
			0.069	0.287	0.169
			0.008	0.275	1.353	1.6

Q, quartz; S, silver.

^aτ̄ = Σ f_i τ_i where f_i = α_i τ_i / Σ α_i τ_i

^b<τ> = Σ α_i τ_i

Our samples are spatially heterogeneous in that the labeled DNA is not localized to a specific distance from the silver surfaces and thus display a range of Γ_m values. Some oligomers may be located on the silver, adjacent to the silver-quartz interface or on the quartz distant from the silver. Nonetheless, it is of interest to estimate the increases in the radiative decay rate, which are consistent with our measurements. Examinations of Equation 1 and Equation 2 show that the radiative decay rate is given by

$$\frac{Q_0}{{\tau }_0}=\mathrm{\Gamma }$$ [Eq. 3]

We calculated apparent values of Γ using the intensities and lifetimes observed on the slides without (○) and with (●) silver. We assumed the quantum yields of Cy3-DNA and Cy5-DNA, in the absence of silver, were the same as we observed previously in solution (17); that is, 0.24 and 0.20, respectively. It is difficult to interpret the intensities on the silver-coated surfaces because the intensities are dependent on both an increased quantum yield of the fluorophores and increased excitation. The latter effect is due to an enhanced electric field near the particles caused by interactions between the free electrons in the metal and the incident light (10). With no consideration of this “lightening rod effect,” the apparent quantum yields of Cy3 and Cy5 exceed unity on the silver surface. Since this is impossible, there must be some component of the enhanced field effect that contributes to the increased intensities in Figure 2. To estimate Γ_m, we assumed the quantum yield on the silver surface increases by 1/Q_o (10), which is the largest possible value and implies that Q_m is unity.

The intensity decays are heterogeneous and described by the multiexponential model

$$I(t)=\Sigma {\alpha }_i\text{exp}(-t/{\tau }_i)$$ [Eq. 4]

where α_i is the amplitude of the component with a lifetime τ_i, and Σ α_i = 1.0. In solution, Cy3-DNA intensity decay can be approximated with two lifetimes τ₁ = 0.38 ns and τ₂ = 1.33 ns with amplitudes α₁ = 0.623 and α₂ = 0.377. For Cy5-DNA, a good fit is achieved with τ₁ = 0.43 ns and τ₂ = 1.44 ns with amplitudes α₁ = 0.471 and α₂ = 0.529. The average lifetimes in solution are slightly shorter than on quartz (Table 1). We believe that this is an effect of more rigid environment on quartz than in solution; that is, some non-radiative deactivation via rotations of the fluorophore is hindered on the surface.

While the lifetimes on the unsilvered area are only slightly nonexponential (11% and 16% for the short components), the lifetimes on the silvered areas are strongly heterogenous. We believe the multiexponential decays represent the range of possible distances between the probe and the surface. The frequency responses shift dramatically to higher frequencies (Figure 3), indicating shortened lifetimes. For the lifetime on the surfaces, we used the amplitude-weighted lifetime <τ> = Σ α_i τ_i. Using these assumptions, we calculated the relative increase in the radiative decay rate Γ_m/Γ. These calculations suggest that the radiative decay rates of Cy3-DNA and Cy5-DNA are increased 40- and 100-fold, respectively, on the silver-coated part of the slides.

It is of interest to consider the increased signal that could be obtained using silver-coated surfaces. The total emission detectable from a fluorophore is usually limited by its photostability (6,20). Photochemical degradation-occurs in the excited state so that a decreased lifetime should allow the fluorophore to undergo more excitation-deexcitation cycles before photo-bleaching. Therefore, we examined the emission intensity of Cy3-DNA and Cy5-DNA with continuous illumination (Figure 4). These traces of intensity versus time were obtained with the same incident intensity so that the time-zero values represent the increased intensities shown in Figure 2. If the time-zero values are normalized then the relative rates at which the emission decreases are about the same on the quartz or silver regions of the slides. This result means that the increased intensity found on the silvered slide is not obtained at the expense of more rapid photobleaching. The relative areas under these curves (I_m/I₀) are 4.8 and 6.5 for Cy3 and Cy5, respectively. In other words, in this experiment, it was possible to extract within 5 min 4.8 and 6.5 times more photons from silvered areas for Cy3-DNA and Cy5-DNA, respectively. Thus, the use of substrates coated with silver particles can result in a substantial increase in the signal observed from cyanine-labeled DNA. In Figure 4, we present photobleaching data with the excitation power adjusted to yield the same emissive photon flux at time zero. These data also show that the number of extracted photons within the time of measurements is higher on the silvered areas than on the quartz alone.

Figure 4.

Photostability of Cy3-DNA and Cy5-DNA on APS-treated slides, with and without silver island films, for the same incident power (left), and the excitation intensity adjusted to yield the same emission intensities on quartz and silver (right).

In this experiment, there was limited control of the distance between the metal and the fluorophores, and the fluorophores were probably present in areas outside the regions where enhanced emission was obtained. It is therefore likely that even larger increases in intensity from labeled DNA can be obtained with localization of the fluorophores at optimal distances from the metallic surfaces.