Optically-Enhanced, Near-IR, Silver Cluster Emission Altered by Single Base Changes in the DNA Template

Abstract

Few-atom silver clusters harbored by DNA are promising fluorophores due to their high molecular brightness along with their long- and short-term photostability. Furthermore, their emission rate can be enhanced when co-illuminated with low-energy light that optically depopulates the fluorescence-limiting dark state. The photophysical basis for this effect is evaluated for two near infrared-emitting clusters. Clusters emitting at ~800 nm form with C₃AC₃AC₃TC₃A and C₃AC₃AC₃GC₃A and both exhibit a trap state with λ_max ~ 840 nm and an absorption cross section of 5–6 × 10⁻¹⁶ cm²/molec that can be optically depopulated. Transient absorption spectra, complemented by fluorescence correlation spectroscopy studies, show that the dark state has an inherent lifetime of 3–4 μs and that absorption from this state is accompanied by photoinduced crossover back to the emissive manifold of states with an action cross section of ~2 × 10⁻¹⁸ cm²/molec. Relative to C₃AC₃AC₃TC₃A, C₃AC₃AC₃GC₃A produces a longer-lived trap state and permits more facile passage back to the emissive manifold. With the C₃AC₃AC₃AC₃G template, a spectrally distinct cluster forms having emission at ~900 nm and its trap state has a ~four-fold shorter lifetime. These studies of optically-gated fluorescence bolster the critical role of the nucleobases on both the formation and excited state dynamics of these highly emissive metallic clusters.

Introduction

Fluorescence imaging opens a mechanistic window on biological processes through its ability to temporally and spatially resolve biochemical events with molecular specificity.^1,2 The development of fluorescent biological labels has yielded access to a broad spectrum of biological events, combining high sensitivity with minimal perturbation. High fluorophore brightness is attainable using lasers to drive electronic transitions, often producing bright integrated and spectrally resolved emission.³ However, emission rate is too-often limited by photoinduced excursions from the fluorescent manifold of states, as evidenced by plateaus in emission intensity and corresponding fluorescence intensity fluctuations.^4–7 This dark-state shelving coupled with relatively long lifetimes in these nonemissive states severely restrict obtainable molecular emission rates. Furthermore, enhanced photobleaching from excited electronic states limits observation times. These detrimental effects can be partially ameliorated using oxidizing and reducing agents to interrupt electron transfer that may sometimes be associated with fluorescence quenching.^8–10 Such additives are designed to suppress triplet populations while also regenerating the chromophore in the singlet electronic state, but their introduction may also perturb the system under study and possibly quench emission. Thus, each chromophore and imaging experiment must be evaluated individually to balance these effects.

Circumventing chemical approaches, populations in dark electronic states can also be optically regulated.^11–15 One approach uses sequential excitation to induce crossover from transiently stable dark states back to the emissive manifold of states via coupling with higher lying electronic levels. Rapid passage back to the emissive manifold thus increases emission intensity by increasing steady-state population of fluorescence-generating molecules. This enhancement depends on the natural lifetime of the dark state and rates of forward and reverse dark state crossings. The power of optically modulating dark state populations is realized when a primary laser produces fluorescence and a secondary laser at a different wavelength depopulates the dark state.¹⁶ For fluorescence imaging, this approach avoids a great deal of the fluorescence background that is inherent in biological systems. Further augmenting the gains from high energy photoswitches,¹⁶ even greater sensitivity enhancements are realized by wavelengths longer than fluorescence for optical depopulation of transient dark states.¹³ As the low energy secondary laser does not increase the fluorescent background, target fluorescence is externally modulated and distinguished from the total background.

DNA-templated silver clusters are a recently developed class of fluorophore that are particularly responsive to optical modulation of their emission. Like their Ag⁺ precursor, reduced silver clusters are stabilized both locally and globally by the electron-rich nucleobases.^17–19 Within oligonucleotide templates, sequence variations yield specific clusters with distinct emission spanning the blue-green to near-infrared spectral region.^20–22 Chemical reduction of oligonucleotide-complexed Ag⁺ typically produces a range of absorbing species, clouding cluster size determinations for the emissive species, unless purification and elemental analysis can be performed.^22,23 These, coupled with size, pH and spectroscopic studies all indicate Ag cluster complexation with the bases of intact ssDNA.^20,24

As bright near infrared emitters are both important and scarce, we have recently focused on silver clusters with near-infrared electronic transitions.²² In parallel with a favorable molecular brightness (product of extinction coefficient and fluorescence quantum yield) of 60,000 M⁻¹ cm⁻¹, especially when compared with spectrally similar emitters,²⁵ this cluster has a short ~μsec dark state lifetime that facilitates both high sustained emission rates relative to more standard dyes such as Cy7 and fluorescence enhancement using dual excitation by short and long wavelength lasers. The present studies investigate the role of subtle base modifications on bright and (optically modulatable) dark state cluster photophysics in a newly identified family of related near-infrared emitting silver clusters. This family is based on the previously identified sequence C₃AC₃AC₃XC₃Y,²² but with varied bases at the X and Y positions to significantly alter bright and dark state photophysics. Spectroscopic investigations of dark state absorption are particularly relevant to the fluorescence enhancement, and, like the ground state absorption, are shown to significantly alter dark state kinetics and energetics. These variations in the primary sequence further establish how the ssDNA template dictates the excited state dynamics of emissive silver clusters. ^13,14

Experimental Methods

Silver nitrate (Aldrich, 99.9999%) and sodium borohydride (Aldrich, 99%) were used as received. Oligonucleotides (Integrated DNA Technologies and Operon) were purified by desalting by the manufacturer and dissolved in sterilized, deionized water (Barnstead Nanopure ultrapure water system). DNA concentrations were determined by absorbance using molar absorptivities based on the nearest-neighbor approximation.²⁶ Silver nanoclusters were synthesized by mixing an oligonucleotide and AgNO₃ solution for 15 sec in 10 mM citrate buffer at pH = 7, then followed by addition of an aqueous NaBH₄ solution. The resulting solution was vigorously shaken for 1 min. Samples were analyzed after reaction for 8 hours. The oligonucleotide concentrations were either 15 or 300 μM with relative molar amounts of 8 Ag⁺ and 4 BH₄⁻. To emphasize the position in the primary sequence that has been modified, the relevant bases are underlined.

Visible absorption spectra were acquired using a Varian Cary 50 Bio UV-Visible and a Shimadzu UV-2401 PC spectrophotometer. Fluorescence spectra were acquired on Jobin Yvon Horiba Fluoromax-3 and Photon Technology International QuantaMaster spectrofluorimeters. Correspondence between the excitation and absorption bands in the near infrared allowed the fluorescence quantum yield to be measured using Cy7 (Φ_f = 0.13) and Indocyanine Green as (Φ_f = 0.04) reference chromophores.^27,28 Reversed-phase ion-pair chromatography was performed with a Shimadzu Prominence HPLC system using a semi-prep Gemini C18 column, which is 50 mm long and 10 mm internal diameter and has particles with an average size of 5 μm and a pore size of 110 Å (Phenomenex). The mobile phase was 35 mM triethylamine acetate at pH = 7 and methanol, with a linear gradient of 10%–30% methanol over 8 min followed by a 3 min hold at 10% methanol at a flow rate of 5 mL/min. Separations were conducted at 25 °C. Absorption and fluorescence measurements of the separated species were made using the SPD-M20A and RF-10XL detectors, respectively. Using a FRC-10A fraction collector, the solutions containing the isolated cluster-DNA conjugate were dehydrated and dissolved in water containing 0.8 M nitric acid. After HPLC purification, the quantities of silver and phosphorous were determined by inductively coupled plasma-atomic emission spectroscopy (Optima 7300 DV, Perkin Elmer). To account for different detection efficiencies, control samples comprised of known relative amounts of Ag⁺ and encapsulating 16-based oligonucleotide were used. To determine the oligonucleotide concentration, the stoichiometry of 15 phosphorous atoms per oligonucleotide was used, which accounts for the absent terminal 3′-phosphate.

Fluorescence correlation spectroscopy studies were conducted using diode laser excitation at 690, 840, and 905 nm with current and temperature control (LTC100, Thorlabs). A 63× 1.2 NA water immersion microscope objective (Zeiss) in a laser epi-illuminated geometry excited the sample and collected the fluorescence. A long-pass dichroic (Semrock) was used to reflect the laser into the back of the objective and transmit the emission, which was spectrally filtered before coupling into a 50-μm diameter fiber located at the image plane of the microscope. The fiber was coupled to actively-quenched single-photon counting avalanche photodiode detectors (SPCMAQR14, Perkin Elmer) using a Hanbury Brown-Twiss setup.²⁹ The resulting TTL signal outputs were cross correlated (Correlator.com) to give an autocorrelation free of afterpulsing artifacts, thereby improving time resolution. To determine the cluster concentrations, dilute solutions of Cy7 and Indocyanine Green were used as reference fluorophores to determine the probe volume dimensions using a 3-D Gaussian model. ^22,30

As described in earlier studies, time resolved nanosecond absorption spectra were conducted as pump-probe experiments.¹⁴ The pump laser was operated at 740 nm with an energy of 700 μJ. The fluorescence lifetime was measured using a mode-locked Ti- Sapphire laser (Coherent Mira) with a pulse-picker for adjusting the repetition rate (ConOptics). TTL signals were processed with a photon counting module (SPC-630, Becker Hickl or PicoHarp E, PicoQuant). Fluorescence decays accounted for convolution of the fluorescence decay with the measured instrument response function.

Results

C₃AC₃AC₃TC₃A

The emission at 810 nm from the Ag₁₀-conjugate formed in C₃AC₃AC₃TC₃A is increased when 905 nm irradiation accompanies the primary excitation at 690 nm.²² Time-resolved transient absorption spectroscopy indicates that this enhancement arises from long-wavelength depopulation of a low yield, long-lived, and non-emissive excited state that is weakly coupled to the emissive manifold of states (Fig. 1 and 2). In the μs time regime, a ground-state bleach with a wavelength minimum that coincides with the λ_max in the absorption spectrum is accompanied by an excited-state absorption with a maximum absorbance change at 833 nm. Direct coupling of this excited dark state with the ground state is indicated by the single exponential relaxation with a time constant of ~3 μs and an isosbestic point at 801 nm between the two bands. Given this two-state conversion, extinction coefficients of the excited and ground states are equal at this common intersection of their spectra.³¹ Based on the previously measured ground state absorption cross section of 7 (± 2) × 10⁻¹⁶ cm² molec⁻¹ at 750 nm²² and accounting for the profiles of the ground state and excited state absorption bands, the excited state absorption cross-section at λ_max = 833 nm is 6 (± 2) × 10⁻¹⁶ cm² molec⁻¹.

Figure 1.

Schematic energy level diagram rationalizing the photophysical behavior of the cluster-DNA conjugates. Fluorescence is derived from cycling through the ground and emissive states using both primary and secondary laser excitation. Inverse transition rates into and out of the dark state define τ_on and τ_off, respectively. Decay from the dark state occurs by both natural decay and optical excitation. The entrance time into the dark state τ_on is the inverse of the product of the dark state quantum yield (φ_D) with the excitation rate (k_exc).

Figure 2.

Time-resolved transient absorption spectra and kinetic decays for the 800 nm emitting silver clusters that form with C₃AC₃AC₃TC₃A (left) and C₃AC₃AC₃GC₃A (right). (Top) Spectra acquired in 2 μs increments after the laser pulse (740 nm, 700 μJ). The blue line is the fit of the ground state bleach and the excited state absorption bands. Discretely constructed spectra at longer wavelengths are indicated by the crosses and support the assignment of the λ_max for the excited state absorption. (Middle) Time evolution of the spectra with isosbestic points at 801 nm (left, C₃AC₃AC₃TC₃A) and at 778 nm (right, C₃AC₃AC₃GC₃A). (Bottom) Kinetic traces show the decay of the excited state populations by monitoring the absorbance at 860 nm (left, C₃AC₃AC₃TC₃A) and 850 nm (right, C₃AC₃AC₃GC₃A). The time constants are 2.6 μs and 6.0 μs, respectively. These time constants do not depend on the absorbance wavelength being monitored.

A fuller picture of the excited state dynamics and dual excitation-based photobrightening is obtained by two-color fluorescence correlation spectroscopy (Fig. 3). Fluorescence fluctuations in the 0.1 – 10 μs time range occur much faster than translational diffusion through the optically-defined probe volume (~400 μs). Thus, autocorrelation analysis of the time-dependent fluorescence changes is resolvable into diffusive (g_D(τ)) and excited state (g_ES(τ)) contributions:¹²

Figure 3.

(Top, left) Correlation function for the cluster associated with C₃AC₃AC₃GC₃A collected at an irradiance of 30 kW/cm₂ at 690 nm. (Top, right) The effect of irradiance at 905 nm (with a fixed irradiance of 30kW/cm² at 690 nm) on the dark state relaxation times for the clusters associated with C₃AC₃AC₃GC₃A (crosses) and C₃AC₃AC₃TC₃A (circles). The inset shows the dependence of the relative enhancement using primary excitation at 690 nm at 30 kW/cm² and varying irradiances at 905 nm for the C₃AC₃AC₃GC₃A (crosses) and C₃AC₃AC₃TC₃A (circles) – based clusters. (Bottom, left) Correlation function for the cluster associated with C₃AC₃AC₃AC₃G collected at an irradiance of 60 kW/cm² at 840 nm. The inset contrasts the spectra for the distinct clusters associated with C₃AC₃AC₃GC₃A (dashed line) using λ_ex = 720 nm and C₃AC₃AC₃AC₃G (solid line) using λ_ex = 840 nm. (Bottom, right) The 840 nm irradiance dependence of the dark state relaxation times for the clusters associated with C₃AC₃AC₃AC₃G.

$$G(\tau )=1+\frac{1}{N}{g}_D(\tau )g_{ES}(\tau )$$

where g_ES(τ) is:

$$g_{ES}(\tau )=1+\frac{F}{1-F}{e}^{-\tau /{\tau }_{ES}}$$

in which F is the fractional occupancy of the dark state and τ_ES is the net correlation time for dark state shelving. In conjunction with 690 nm excitation within the emissive manifold of states, secondary excitation at 905 nm selectively excites the dark state. Based on the time-resolved absorption spectra, absorption at 905 nm is expected to produce significant excitation of the dark state population with minimal ground state excitation (Fig. 2). This selective photoexcitation is evident by decomposing the observed correlation times (τ_ES) into entrance times to (τ_on) and exit times from (τ_off) the dark state as a function of secondary laser irradiance (Table 1).³² The τ_on values are not greatly perturbed by the secondary laser, consistent with its energy being off-resonance relative to ground state excitation. In contrast, τ_off values decrease with irradiance at the longer wavelength, supporting selective dark state depopulation. Mechanistic perspective is provided using the following reaction scheme:¹²

Table 1.

Comparison of measured τ_on and τ_off for the dark state associated with the near-infrared emitting clusters that form with C₃AC₃AC₃TC₃A and C₃AC₃AC₃GC₃A. Primary excitation at 690 nm has fixed irradiance (30 kW/cm²)^a and secondary excitation is at 905 nm with varying irradiance.

C3AC3AC3TC3A			C3AC3AC3GC3A
Secondary laser irradiance (kW/cm2)a	τon (μs)	τoff (μs)	Secondary laser irradiance (kW/cm2) a	τon (μs)	τoff (μs)
0	2.7	3.4	0	1.9	4.1
3	2.0	3.0	4	1.9	3.6
8	1.8	2.7	10	1.9	3.0
13	1.8	2.9	17	1.9	2.6
21	1.8	2.2	25	1.9	2.0
30	1.9	2.1	33	1.5	1.6
43	2.1	1.7	39	1.8	1.7

^aIrradiance values derived from the average intensity in the FCS focal volume.

$$\text{Emissive}\text{Manifold}\underset{{\text{k}}_{\text{DS}}}{\overset{{\text{k}}_{\text{SD}}}{\rightleftharpoons }}\text{Dark}\text{Manifold}$$

In this model, the emissive state is strongly coupled with the ground state, and the dark state is weakly coupled to the emissive manifold as described by the relatively small rate constants k_SD and k_DS for entering into and exiting from the dark state, respectively (Fig. 3). Assuming identical diffusion coefficients associated with these two states, a kinetic analysis relates the net correlation time (τ_ES) and the steady-state fractional occupancy (F) to the rate constants:¹²

$$\begin{array}{l}{k}_{DS}=\frac{1-F}{{\tau }_{ES}}\\ k_{DS}=k_{DS}^o+\sigma \gamma I\end{array}$$

where $k_{DS}^o$ is the inherent dark state decay rate constant due solely to nonradiative relaxation, σ(cm²/molec) is the action cross section for reverse intersystem crossing back to the singlet manifold, γ (photon/J) is (hν)⁻¹, and I is the secondary laser irradiance (W/cm²). Based on how the observed rate constant varies with a ten-fold variation in the irradiance of the secondary beam, an inherent decay rate constant of 3.0 (±0.1) × 10⁵ s⁻¹ with a corresponding time constant of 3.3 μs is obtained (Fig. 3). The slope provides the action cross section for reverse crossing, and the value of 0.015 (± 0.002) × 10⁻¹⁶ cm²/molec is comparable to values previously measured for organic chromophores.¹²

C₃AC₃AC₃GC₃A

Within the general sequence C₃AC₃AC₃XC₃A, the three remaining oligonucleotides with X = G, C, and A were investigated. Relative to the thymine derivative discussed above, these modifications produce clusters with similar absorption and fluorescence spectra, thus indicating that essentially the same silver cluster forms with all these templates. Because the oligonucleotides containing thymine and guanine are distinguished by their chemical stability, the subsequent discussion focuses on the photochemical and chemical characteristics of the cluster that forms with C₃AC₃AC₃GC₃A that emits at 770 nm (Fig. 3). As with C₃AC₃AC₃TC₃A,²² a mixture of species forms with C₃AC₃AC₃GC₃A, but these can be separated by reversed-phase HPLC (Fig. 4). Of the three primary species, the near-infrared emitting species is identified by its fluorescence response and associated absorption feature at 720 nm. The absorption spectrum shows that the ubiquitous absorbance at ~400 nm in as-synthesized samples is removed through purification, with only the near IR cluster and UV DNA absorptions remaining. Using atomic emission spectroscopy to quantify the amounts of phosphorous and silver, the known phosphate content of the oligonucleotide was used to determine a relative stoichiometry of 10.0 ± 1.7 Ag:C₃AC₃AC₃GC₃A, which is similar to the value of 9.6 ± 0.8 Ag previously reported for similarly emitting clusters in C₃AC₃AC₃TC₃A ²². Relative to C₃AC₃AC₃TC₃A, the guanine containing oligonucleotide encapsulates a cluster with a similar fluorescence quantum yield of 30 ± 5% and a comparable lifetime of 2.2 ns. Despite these stoichiometry and photophysical similarities, the cluster environment is influenced by nucleobase substitution, as illustrated by the absorption cross section. Identified by its fluorescence signature, the cluster concentration was quantified using the amplitude of the diffusive component of the fluorescence correlation function, and the optically defined probe volume was calibrated using solutions of Cy7 concentration standards.³³ Cluster occupancy varied linearly with dilution, thus enabling extrapolation to the concentrations used for absorbance. The resulting absorption cross section of 3.2 (± 0.4) × 10⁻¹⁶ cm² molec⁻¹ at 720 nm is only about 50% of that measured for the C₃AC₃AC₃TC₃A-templated cluster, but the resulting molecular brightness (Σ×Φ_fl) of 24,000 M⁻¹ cm⁻¹ again compares favorably to chromophores in this spectral region.²⁵

Figure 4.

(Top left) Chromatograms derived using the absorbance at 260 nm (left axis) and fluorescence intensity (cps) using λ_ex = 720 nm and λ_em = 800 nm (right axis) for the cluster-C₃AC₃AC₃GC₃A conjugates. The near infrared emissive species is associated with Peak II. (Top right) Spectra associated with peaks in the chromatogram on the left. (Bottom left) Chromatograms derived using the absorbance at 260 nm (left axis) and fluorescence intensity (cps) using λ_ex = 840 nm and λ_em = 870 nm (right axis) for the cluster-C₃AC₃AC₃AC₃G conjugates. The near infrared-emissive species is associated with Peak II. (Bottom right) Absorption spectra associated with peaks in the chromatogram on the left. The inset expands the view for the spectrum of the near infrared emitter. Note that the spectral capabilities of the instrument absorption do not extend beyond 800 nm.

Further effects of the cluster environment are evident in the excited state dynamics. Time-resolved absorption and fluorescence correlation spectroscopy studies show that the cluster associated with C₃AC₃AC₃TC₃A and C₃AC₃AC₃GC₃A exhibit similar dark state characteristics, but branching out of this state is affected by base identity at the 12^th nucleotide. Spectral decomposition of the time-resolved absorption data identifies the ground state bleach and an excited state absorption (Fig. 2). The λ_max = 841 nm for the dark state absorption feature is similar to that observed for clusters within C₃AC₃AC₃TC₃A. Two-state relaxation is again supported by monoexponential decay from the excited state and by an isosbestic point having ΔOD = 0 at 778 nm between the bleach and absorption bands, and the latter observation yields an excited state absorption cross section of 4.8 (±0.5) × 10⁻¹⁶ cm²/molec at 841 nm. Escape efficiency from the dark state depends on the DNA sequence, as shown by fluorescence correlation spectroscopy (Fig. 3). Using the two-laser excitation scheme with varying excitation rates at 905 nm in conjunction with fixed intensity primary excitation at 690 nm for the C₃AC₃AC₃GC₃Aencapsulated silver cluster, we find that τ_on values are not altered by a ten-fold increase in the irradiance at 905 nm while the τ_off values decrease (Table 1). By analyzing the irradiance dependence of observed decay rate constants, an inherent relaxation rate constant of 2.4 (± 0.2) × 10⁵ s⁻¹ with an associated time constant of 4.2 μs is measured, and this rate is ~20% slower relative to that measured for clusters in C₃AC₃AC₃TC₃A. The resulting action cross section for reverse intersystem crossing is 0.022 (± 0.002) × 10⁻¹⁶ cm² molec⁻¹, which is ~50% higher than that formed in C₃AC₃AC₃TC₃A. This cluster species is also distinguished by its larger relative enhancement derived from τ_off and τ_on. Based on a three-level system with both natural and photoinduced loss from the dark state, the enhancement achieved using dual vs. single laser excitation in the limit of low excitation rate is:

$$\text{Relative}\text{Enhancement}=\frac{I_F^{\text{Dual}\text{Laser}}-I_F^{\text{Single}\text{Laser}}}{I_F^{\text{Single}\text{Laser}}}=\frac{{\tau }_{\mathit{off}}^0-{\tau }_{\mathit{off}}}{{\tau }_{on}+{\tau }_{\mathit{off}}}$$

where $I_{\text{F}}^{\text{Dual}\text{Laser}}$ and $I_{\text{F}}^{\text{Single}\text{Laser}}$ are the fluorescence intensities using dual and single laser excitation, respectively and ${\tau }_{\mathit{off}}^0=1/k_{DS}^0$ represents the time for strictly natural decay from the dark state. The comparable on times but longer off times of clusters in C₃AC₃AC₃GC₃A relative to those in C₃AC₃AC₃TC₃A suggests a cluster environment that is more amenable to fluorescence brightening by the secondary laser (Fig. 3 and Table 1).

C₃AC₃AC₃AC₃G

A primary feature of DNA-templated silver clusters is the profound effect of base sequence on cluster type.²⁰ Further consideration of sequence effects shows that the ~800 nm emitter also forms with X = A, T, and C in C₃AC₃AC₃AC₃X, yet a distinct species emitting at 900 nm forms with X = G (Fig. 3). With this template, four major species are chromatographically separated, and the new cluster is identified by its spectral signatures (Fig. 4). Atomic emission studies following chromatographic isolation gives a stoichiometry of 9.1 ± 0.5 Ag:C₃AC₃AC₃AC₃G. Thus, the 800-nm emitters are both consistent with 9 or 10 Ag atom clusters, while the 900-nm emitter is only consistent with a 9-Ag atom species. Beyond the spectral distinctions, the cluster that forms with C₃AC₃AC₃AC₃G is further distinguished by its emissive and dark state photophysics. Relative to the ~800 nm emitter that forms with C₃AC₃AC₃GC₃A, the ~900 nm emitter has a five-fold lower fluorescence quantum yield (6 ± 2 %) and a five-fold higher absorption cross section (16.8 (± 1.1) × 10⁻¹⁶ cm² molec⁻¹ at λ_max = 840 nm). Together, the resulting molecular brightness of 26,000 M⁻¹ cm⁻¹ and short fluorescence lifetime of 1.4 ns again make this a promising chromophore to complement the limited choices of chromophores in this spectral region.²⁵ Fluorescence correlation studies show weak coupling to a dark state, but the rate of relaxation is faster for the ~900 nm emitter relative to those of the ~800 nm emitters (Fig. 3 and Tables 1 and 2). Irradiance measurements at 840 nm show an inherent decay rate constant of 9.4 (± 1.7) × 10⁵ s⁻¹ with a corresponding time constant of 1.1 μs. Evidence of an optical contribution to the excited dark state dynamics, however, is provided by the τ_on and τ_off times (Table 2). Coupled with the greater primary laser-induced excitation rate, the probability of entering the dark state increases at higher irradiances. In tandem, the exit rates increase with primary laser irradiance, suggesting that primary excitation laser also accelerates dark state depletion. Efforts to spectroscopically identify this state were hampered by the ground state and dark state spectral overlap, as indicated by the τ_on and τ_off values for single laser excitation (Table 2).

Table 2.

Comparison of measured τ_on and τ_off for the dark state associated with the near-infrared emitting clusters that form with C₃AC₃AC₃AC₃G. Primary excitation at 840 nm is varied with no secondary illumination.

Primary laser irradiance (kW/cm2)	τon (μs)	τoff (μs)
6	17.5	2.4
9	8.7	1.5
14	2.3	0.8
23	1.6	0.8
55	0.9	0.6
57	1.0	0.6
102	0.8	0.5
132	0.7	0.4
138	0.4	0.4
187	0.5	0.3
267	0.4	0.3
287	0.5	0.3

Discussion

A key observation from our studies is that dark electronic states in near infrared emitting silver clusters can be optically manipulated to enhance fluorescence through increasing steady-state population within the fluorescent manifold of states.^11–13 Optically induced excited state depopulation allows these relaxation pathways to be deciphered, while simultaneously demonstrating that emission is enhanced by repopulating the emissive manifold. Spurred by the desire to enhance label visibility for fluorescence imaging, a broad study of chromophores revealed that the degree of reverse intersystem crossing and hence fluorescence enhancement is specific to the dye, to its wavelength for excitation out of the triplet state, and to its environment.¹² With most of these systems, effective cross sections for reverse intersystem crossing of 10⁻¹⁹ - 10⁻¹⁸ cm²/molec are derived from kinetic modeling.^12,34

Dark electronic states in DNA-conjugated silver clusters can also be optically addressed.^13,14,22 For clusters that emit at ~700 nm, a charge-transfer state is supported by a large change in the dipole moment, and cytosine anion-suggestive transient absorption spectra.¹⁴ Our time-resolved absorption spectra demonstrate that near infrared emitting silver clusters conjugated with C₃AC₃AC₃TC₃A and C₃AC₃AC₃GC₃A exhibit similar dark electronic state characteristics with λ_max ~ 840 nm, and with absorption cross sections of 5–6 × 10⁻¹⁶ cm²/molecule. Photophysically distinct from the shorter wavelength emitters studied previously, the dark states of the longer-wavelength emitting, 9 – 10 atom clusters relax with time constants of 3 – 4 μs, and can be optically promoted using a second, lower energy laser. The net emission rate is enhanced because the steady state population within the fluorescent state manifold is increased due to optical depopulation of the transient dark state, having an action cross section of ~10⁻¹⁸ cm²/molec at 905 nm. This value is comparable to reverse intersystem crossing cross sections exhibited by organic fluorophores, thereby supporting efficient enhancement of emission via optical methods.¹² Insight into the nature of the dark electronic state is provided by comparison with previously studied red emitters which have a broad transient absorption band centered at ~650 nm ascribed to electron transfer from the cluster to the cytosine nucleobase.¹⁴ The ~800 nm emitters are distinguished by their relatively narrow transient absorption feature at ~840 nm that may be enveloped by the broader transition observed for the ~700 nm emitters. For both emitters, dark state excitation energies could feasibly promote oxidation of these small clusters.^14,35

In addition to demonstrating long-wavelength optically enhanced near IR fluorescence, another key observation is the impact of the nucleobase environment. Comparing the clusters that form in C₃AC₃AC₃TC₃A and C₃AC₃AC₃GC₃A, the guanine derivative has a longer transient lifetime within and a higher action cross section for reverse crossing out of the dark state, thereby leading to its greater degree of fluorescence enhancement. Furthermore, primary sequence effects on the types of clusters that form also extend to the near infrared, as demonstrated by the formation of a new near infrared emitter with emission at 900 nm using C₃AC₃AC₃AC₃G. Not only is the energy spacing in the emissive states influenced, but the excited state dynamics are distinguished by the short lifetime. These observations suggest that both stabilization of particular clusters and their excited state properties are dictated by the nucleobase interactions that direct cluster formation.

Conclusion

The primary conclusion from our steady state, transient absorption, and fluorescence correlation spectroscopy studies is that nucleobase sequence dictates the types of templated clusters and their excited state dynamics. Within the general sequence C₃AC₃AC₃XC₃Y, the DNA-templated clusters exhibit emissions at ~800 and ~900 nm with (X, Y) = (T/G, A) and (X, Y) = (A, G), respectively. The excited electronic states of these emissive metallic clusters couple to trap states, and lifetimes in these states also depend on the base sequence of the encapsulating oligonucleotide. Complementing this ability to optically gate emission using long wavelength excitation out of a transiently populated dark state, these fluorophores also display relatively high molecular brightness (>10⁴ M⁻¹cm⁻¹) in the biologically important near infrared spectral range. These characteristics portend the promise of DNA-templated silver clusters as fluorescent labels.