1.

Introduction

Quantum dots (QDs) are luminescent semiconductor nanocrystals that, due to the quantum confinement effect and macromolecular size, combine unique photophysical features such as broad strong absorption and size-dependent efficient photoluminescence (PL).^1,2 The mobility in liquid phases and high surface-to-volume ratio as well as effective methods established for their bioconjugation make QDs promising objects for bioanalytical techniques including in vitro diagnostics and PL imaging.³ Noticeably, QDs have a reduced tendency for PL self-quenching and photobleaching, which are common drawbacks observed for organic dyes or fluorescent proteins.⁴ This improved resistance against photodegradation effects also enables high sensitivity and usability in high-throughput molecular diagnostics.⁵ Over the last two decades, QDs have attracted considerable attention as universal luminescent nano-objects for numerous fields including medicine,⁶ pharmacology,⁷ and chemical sensing.⁸ In recent years, various QDs have been used in biological and biomedical applications, such as multiplexed systems for DNA detection,⁹ biomarker sensing,^10,11 luminescent immunoassays,^12,13 bioassays,¹⁴ optical coding,^15,16 protein concentration determination,¹⁷ drug tracking,¹⁸ single molecule tracking,¹⁹ photodynamic therapy,²⁰ and intracellular imaging.^21,22 Continuously, various new QD-based nanosystems are appearing, and new application fields are growing dynamically.

In recent years, numerous systems have been presented in which QDs are combined with organic fluorescent moieties to achieve additional properties beyond those provided solely by QDs. In this direction, a wide variety of systems can be found, including a combination of nanorods and dyes,²³ pyrene-functionalized nanoparticles to detect nitroaromatic compounds such as nitroanilines and nitrobenzenes,²⁴ CdSe QDs functionalized with a naphthalimide dye to yield unconventional QD quenching,²⁵ graphene QDs combined with europium ions to recognize phosphates,²⁶ QD-carbon nanotube conjugates for photoacoustic and PL detection of circulating cells with flow cytometry in vivo,²⁷ doxorubicin-QD conjugates for photo-controlled drug delivery,²⁸ and QD-europium ion conjugates to detect nucleoside triphosphates.²⁹ A QD-Lucigenin conjugate sensor has also been developed to sense chloride anions in an aqueous environment with a sensitivity down to around 300 nM.³⁰ Interesting examples of luminescent QD-based conjugates presented a cascade FRET from a conjugated polymer to a QD and then to an organic dye,³¹ or multistep FRET from a Tb-complex to a QD and then to an organic dye.^10,32,33 Very recently, CdTe QDs functionalized with a naphthyridine dye were developed to detect guanosine nucleotides in an aqueous environment.³⁴ Due to specific and selective naphthyridine-nucleotide interactions combined with the optical properties of the QDs, those conjugates showed improved sensitivity in comparison to naphthyridine or polystyrene nanoparticle-based sensors.³⁵ These nanosensors were able to detect cyclic 3’5’ guanosine monophosphate (cGMP) down to $70\mathrm{ng}/\mathrm{mL}$.

In numerous instances of the above summarized QD-dye systems, Förster resonance energy transfer (FRET) was the main mechanism responsible for the specific signal modulation. FRET is a nonradiative energy transfer from one emitting moiety (donor) to another energy-receiving moiety (acceptor). In fundamental considerations, FRET is a dipole-dipole interaction and it requires both close donor-acceptor spatial proximity (ca. 1 to 20 nm) and spectral overlap between donor emission and acceptor absorption.³⁶ Lanthanide complexes as FRET donors provide many advantages, which are mainly related to their long excited-state decay times, large effective Stokes’ shifts, and well-defined and narrow emission peaks.^37,38 The use of lanthanide complexes in FRET applications facilitates the transfer of their long PL decay times to acceptors, which leads to long-living species emitting at wavelengths different from those of the lanthanide ion. Such systems have been demonstrated with simultaneous FRET to five different acceptors using organic dyes³⁹ or QDs.⁴⁰ Because QDs have very narrow emission bands and large absorption cross sections over a broad wavelength range, they are superior FRET acceptors in combination with lanthanides.^41–43 The lanthanide-to-QD FRET was demonstrated for the first time with biotin-streptavidin-based binding systems using Eu- and Tb-based complexes.^44,45 Although there exist FRET systems that combine lanthanides and QDs, they have always been based on FRET that was enabled by biological recognition (e.g., biotin-streptavidin). A direct integration of lanthanides and QDs within one nanoparticle could overcome the necessity of binding-established FRET and yield a single lanthanide-QD emitter. Apart from medical applications, such QD-based nanoconjugates could find their application in display technologies, where long-living emissions at freely selected and well-defined wavelengths are highly desirable properties.

In this contribution, we present a QD-based conjugate architecture which combines an emission spectrum typical of a QD with a decay time typical of a europium complex. The preparation strategy, photophysical characterization, and the performance in biotin-streptavidin PL assays are also described. The QD-EuC-biotin conjugates have a high potential to be applied to, but not limited to, routine applications in fluoroimmunoassays and cellular imaging.

2.

Materials and Methods

3.

Results and Discussions

Hydrophobically capped QDs were transferred into aqueous solution using overcoating with an amphiphilic polymer, which has advantages and disadvantages compared to standard ligand–exchange protocols.^22,48,49 Depending on the method, several properties can be varied: the distance between attached ligands and the QD surface, nonspecific ligand-QD surface interaction, and colloidal stability.^48,49 Prior to the practical applications, this consideration should be taken into account and the method to transfer initially hydrophobic QDs into aqueous solution needs to be selected based upon the particular requirements of the application. In our present case, we desired QDs with a high colloidal stability and thus opted for coating them with an amphiphilic polymer.⁴⁶

The polymer prepared within this study and its arrangement on a QD is shown in Fig. 1. The octyl chains of trioctylphosphine oxide (TOPO), which are present on the hydrophobic QDs as supplied by the vendor, intercalate in organic solvent with the dodecyl chains of the amphiphilic polymer (by hydrophobic interaction) to form a hydrophobic buffering shell around the QD.⁴⁶ Importantly, the amphiphilic polymer backbone is built of maleic anhydride rings. When placed in aqueous solution, the maleic anhydrides hydrolyze, which results in a large number of carboxylic moieties on the QD surface to provide the QDs with colloidal stability.

Fig. 1

The assembly of the functional polymer on the QD surface. The QD (either QD655 or QD705) is surrounded by trioctylphosphine oxide chains (coming from the QD preparation) assembled with the dodecyl chains of the amphiphilic polymer. Biotin (to facilitate binding to streptavidin) and EuC (to pump QD with long-living emission) were incorporated into the polymer prior to its assembly on the QD. The remaining carboxylic moieties facilitate stability and solubility of such construct in aqueous solutions.

The maleic anhydride rings also facilitate uncomplicated coupling of amine-containing molecules to the polymer through amide bond formation, which in the present cases has been used to premodify the polymer (before coating the QDs) with EuC-PEG-biotin.⁴⁶ After transfer of the polymer-coated QDs to an aqueous solution, the carboxylic moieties would also facilitate an uncomplicated polymer postfunctionalization in an aqueous environment, e.g., via 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) coupling. As seen in Fig. 2, after the preparation process, the EuC-polymer emission spectra remained in good agreement with the spectra typical for EuC. Only the relative emission intensities have changed reflecting the change in the immediate vicinity of the complex. Additionally, the EuC-polymer excitation spectrum was shifted about 15 nm to the red.

Fig. 2

Steady-state PL excitation and emission spectra taken for EuC free in solution and when conjugated to the amphiphilic polymer.

In our system, efficient FRET is observed due to close spatial proximity between EuC donors in the polymer shell and the QD acceptors. A distance estimation based on assuming a separation of the EuC from the inorganic QD surface by $\sim 30$ single C-C bonds (each with a length of 0.12–0.15 nm) yields a distance of 3–5 nm, assuming the dodecyl chains in their extended form. For calculation of FRET efficiencies, one also needs to take into account the distance from the QDs surface to its center corresponding to the dipole center. Using the core/shell radius $r_c(=d_c/2)$, which is around $r_c=3–5\mathrm{nm}$ for both kinds of QDs, the estimated EuC-QD donor–acceptor distance is approximately $r=6–10\mathrm{nm}$. With a Förster radius of $R_0=10\mathrm{nm}$ for the EuC-QD655 system,¹⁷ we calculated a FRET-efficiency $[{\eta }_{\mathrm{FRET}}=(R_0^6)/(r_6+R_0^6)]$ of ${\eta }_{\mathrm{FRET}}=0.5–0.96$. Notice that the Förster radius is a theoretical value, which depends on (1) donor quantum yield in the absence of an acceptor, (2) overlap between the donor’s emission and the acceptor’s absorption, (3) the orientation factor between both dipoles, and (4) the refractive index of the medium.

Time-gated intensity detection in time-resolved spectroscopy allows for minimizing the undesired background effects resulting from directly excited QDs. Under ultraviolet (UV) light excitation at 337 nm, the QDs are excited both directly and indirectly through FRET. In a detection timeframe starting from 100 μs, PL coming from directly excited QDs is negligible and the emission coming from FRET-sensitized QDs can mainly be observed. Regarding the extinction coefficients of EuC [$\epsilon (\mathrm{EuC})=\mathrm{31,000}{\mathrm{M}}^{-1}{\mathrm{cm}}^{-1}$ ⁵⁰] and the QDs [$\epsilon (\mathrm{QD}655)=\mathrm{11,500,000}{\mathrm{M}}^{-1}{\mathrm{cm}}^{-1}$ and $\epsilon (\mathrm{QD}705)=\mathrm{17,000,000}{\mathrm{M}}^{-1}{\mathrm{cm}}^{-1}$ [at 337 nm, which presents a ca. 350-to-550-fold larger absorptivity for the QDs, the steady-state contribution of EuC to the emission spectra of EuC-QD samples can be neglected and therefore an interpretation of steady-state spectra for EuC-to-QD FRET is not conclusive. For time-resolved measurements, we used a bandpass filter, which passes light at a QD-characteristic (Eu-characteristic) wavelength and blocks most of the contributions from Eu-emission (QD-emission). In our case, the crosstalk of the time-gated EuC emission intensity to the acceptor channel was found to be 5% (at $655\pm 13\mathrm{nm}$) and 1% (at $740\pm 13\mathrm{nm}$) of the intensity value collected in the donor channel at $(620\pm 10)\mathrm{nm}$.

In Fig. 3, steady-state PL measurements are presented for the QD-EuC nanocomplexes. In those spectra, residual emission in the 400–500 nm range can be observed, which we attributed to emission of the terpyridine present in the EuC organic antenna. The terpyridine emission is most probably due to an extraction of ${\mathrm{Eu}}^{3+}$ ions from the chelating antenna during purification using gel electrophoresis. This assumption is supported by emission spectra of the EuC-functionalized polymer before coating to the QD (Fig. 2), which did not show any terpyridine emission. In order to “recharge” the antennas with ${\mathrm{Eu}}^{3+}$ ions, we added 10 μL of 25 mM ${\mathrm{EuCl}}_3$ (in 50 mM citric acid, $\mathrm{pH}=4$) to the 120 μL solution of 50 nM QD-EuC. Adding a large excess of ${\mathrm{EuCl}}_3$ allowed the chelates that have lost their ions to coordinate new ${\mathrm{Eu}}^{3+}$ ions in order to guarantee a maximum chelate recovery for efficient FRET to the QDs. The resulting spectra gave evidence of two main effects. First, the addition of ${\mathrm{EuCl}}_3$ led to a strong reduction in the terpyridine emission, thus indicating the successful ${\mathrm{Eu}}^{3+}$ coordination into the chelating antenna. Compared to the QD655-EuC samples [Fig. 3(a)], the suppression of terpyridine emission becomes much more evident for the QD705-EuC samples [Fig. 3(b)]. This effect can be associated with different amounts of ${\mathrm{Eu}}^{3+}$ ions extracted from the terpyridine chelating antennas for those two systems. In parallel to the decrease of antenna emission, the QDs emission intensity is strongly increased (fivefold for QD655 and threefold for QD705), which suggests that these two processes are related to each other by the improved EuC-to-QD FRET. However, taking into account the large difference in molar absorptivities of EuC and the QDs (vide supra), the strong QD PL intensity increases cannot be purely caused by FRET from EuC to QD. We assume that this effect is a mixture of QD-stabilization (increasing the steady-state PL intensity of directly excited QDs) and increased FRET-sensitization by EuC.

Fig. 3

Luminescence emission spectra (${\lambda }_{\mathrm{ex}}=337\mathrm{nm}$) of QD655-EuC-biotin (a) and QD705-EuC-biotin (b) nanoconjugates before (blue) and after addition of 10 μL (black) of 25 mM ${\mathrm{EuCl}}_3$ in 50 mM citric acid, $\mathrm{pH}=4$.

In order to verify that the addition of ${\mathrm{EuCl}}_3$ leads to efficient FRET from EuC to QD (and not only better QD emitters), we performed time-resolved PL measurements. The PL decays for QD655, EuC, and QD655-EuC-biotin are shown in Fig. 4.

Fig. 4

(a) PL decays collected for QD655 without EuC and biotin (black curve), QD655-Eu-biotin before (green curve) and after (blue curve) the addition of 10 μL of 25 mM ${\mathrm{EuCl}}_3$ in 50 mM citric acid, $\mathrm{pH}=4$. (${\lambda }_{\mathrm{exc}}=337\mathrm{nm}$, emission bandpass filter – $(665\pm 13)\mathrm{nm}$). The red curves correspond to fits with a triexponential function. The different time zones indicated on top of the graph represent the different PL intensity contributions of detector saturation D (instrument response), FRET, and pure (unquenched) EuC PL. (b) Control experiments showing the decays collected for pure PBS buffer (red curve), 3 mM ${\mathrm{EuCl}}_3$ in PBS (green curve), and the EuC-biotin polymer in PBS (black curve). Note: The EuC-biotin decay is shown to illustrate the long decay of pure EuC. As it was collected for a pure EuC polymer solution (without attachment to QD655 and subsequent separation of pure EuC), it had a significantly higher concentration and intensity than the pure EuC PL presented in Fig. 4(a).

Similarly to the steady-state measurements, in time-resolved measurements an increase in QD PL is also observed upon ${\mathrm{EuCl}}_3$ addition to the nanoconjugate solution. However, PL decay curves (taken over a 6 ms timeframe) clearly show QD-FRET sensitization by long-living EuC. The black curve in Fig. 4(a) contains a microsecond decay component which is caused by strong direct excitation of the QDs and the buffer by the nitrogen laser. Although QD decay after direct excitation only lasts several microseconds (decay times with tens to hundreds of nanoseconds), the decay in Fig. 4 lasts up to $<1\mathrm{ms}$. This decay is caused by saturation of the detection setup directly after intense-pulsed nitrogen laser-excitation [which can even be found for pure PBS buffer excitation, cf. red curve in Fig. 4(b)] and thus mainly reflects the instrumental response. The green curve in Fig. 4(a) contains several decay components:

Addition of ${\mathrm{EuCl}}_3$ leads to a strong intensity increase of the FRET component, whereas the pure EuC PL and detector saturation components remain unchanged [blue curve in Fig. 4(a)]. The decay curve of pure ${\mathrm{EuCl}}_3$ in solution [green curve in Fig. 4(b)] is only slightly more intense than and has a similar time range to that of the buffer decay [red curve in Fig. 4(b)]. Therefore, the intensity increase from the green to the blue curves in Figure 4(a) cannot be attributed to PL originating from pure ${\mathrm{Eu}}^{3+}$ ions in the sample solution. In order to quantify the single contributions, we fitted the decay curves with a triexponential function Eq. (1), which takes into account detector saturation (${\tau }_1$), pure EuC PL (${\tau }_2$), and FRET (${\tau }_3$).⁵¹

Here $A_1$, $A_2$, and $A_3$ are the amplitude fractions (with $A_1+A_2+A_3=1$). Because the short decay component is caused by the instrument response to a very strong short-lived signal, this component cannot be taken into account for a quantitative evaluation of direct QD excitation. The QDs are very strongly excited at 337 nm and therefore emit a very strong PL signal. As mentioned above, this strong signal saturates the photomultiplier detectors; therefore, the observed decay represents a detection setup decay signal of the photomultiplier electronics rather than the optical decay of the QDs. However, fitting for this component allows a subtraction of this unwanted background and a re-evaluation and quantification of the unquenched EuC and the FRET-sensitized QD components using Eq. (2):

Table 1

The decay times τi and corresponding amplitudes Ai and Bi obtained from fitting PL decays in Fig. 4 using Eqs. (1) and (2). QD655-EuC-biotin (1) corresponds to the sample resulting from nanobioconjugate preparation and QD655-EuC-biotin (2) to the sample after addition 10 μL of EuCl3.

Based on the average FRET-decay time value ($\langle {\tau }_3\rangle =230\pm 20\mu \text{s}$) and the average decay time of unquenched EuC ($\langle {\tau }_2\rangle =900\pm 130\mu \text{s}$), the FRET efficiency was calculated to be ${\eta }_{\mathrm{FRET}}=1-\langle {\tau }_3\rangle /\langle {\tau }_2\rangle =0.74\pm 0.6$. Using the Förster radius of $R_0=10\mathrm{nm}$, this leads to an average EuC-QD donor–acceptor distance of $r=R_0{[\langle {\tau }_3\rangle /(\langle {\tau }_2\rangle -\langle {\tau }_3\rangle )]}^{1/6}=8.4\pm 0.5\mathrm{nm}$, which is in very good agreement with the estimated distance range from 6 to 10 nm (vide supra) and with previously determined values for the organic polymeric shell thickness.⁴⁷

3.1.

Photoluminescence Bioassay

The biotin–streptavidin recognition system is often used in biological recognition assays due to having the highest binding affinity constant (ca. ${10}^{-14}\mathrm{mol}/\mathrm{L}$) known for noncovalent interactions. Due to the strong biotin-streptavidin interaction, biotin is a convenient chemical group to test the biosensing usability of our complexes in an uncomplicated way. In our study, the PEG-biotin was attached to the amphiphilic polymer to provide biotin functionality to the QD655-EuC-biotin. Microtiter plate wells covered with different streptavidin concentrations (1, 2, $3\mu \text{g}/\mathrm{mL}$) were exposed to 50 nM QD655-EuC-biotin solutions in PBS. After incubation with QD655-EuC-biotin (and QD655-biotin as a control) and subsequent washing, the PL assays were performed using time-gated (200–2000 μs) PL intensity detection (20–200 ns time gate for QD655-biotin). The assay results are presented in Fig. 5. One can observe an enhancement in the PL signal with increasing streptavidin concentration. Notably, for the presented system, we could observe a 30-fold enhancement in PL from 0 to $2\mu \text{g}/\mathrm{ml}$ (0–38 nM) streptavidin, which is remarkable in comparison to the situation when only QDs are used. We note that the detection setup for the QD655 biotin is not optimal because time gating and UV excitation at low repetition rates would not be necessary. However, in order to be able to use the same experimental conditions, we also used time gating in a relevant time range for QD PL emission. Even without this comparison, our results demonstrate a highly sensitive detection of biotin–streptavidin binding, and therefore, the potential of our nanobioconjugates for their use in PL bioassays such as heterogeneous immunoassays. For example, in a heterogeneous assay (washing after each step is omitted in the following description) a primary “catcher” antibody is adsorbed on the well surface. After passivation, the catcher antibodies are exposed to a sample containing an analyte, i.e., an antigen. Then, a secondary antibody (e.g., conjugated with a fluorescent dye) – a “detector” – is added for quantification. Replacing the fluorescent dye on the detector antibody with streptavidin would allow the attachment of our biotinylated nanobioconjugate and a sensitive time-gated detection without autofluorescence background. Using different QD colors (e.g., QD655 and QD705 as in our study) would allow for multiplexed time-gated detection. Apart from the use in immunoassay, such constructs can find their application in other fields such as imaging or display technologies.

Fig. 5

PL signal (integrated over 1800 μs (200–2000 μs) of QD655-EuC-biotin and commercial biotinylated QD655 [integrated over 180 ns (20–200 ns)] collected in microtiter plate wells covered with 1, 2, and $3\mu \text{g}/\mathrm{mL}$ of streptavidin. The signals were normalized to 1 corresponding to $0\mu \text{g}/\mathrm{mL}$ streptavidin. A well only with BSA without streptavidin ($0\mu \text{g}/\mathrm{mL}$) has been added to test nonspecific interactions. The control corresponds to a signal from a well in which no BSA and no streptavidin was present.

4.

Conclusions

In this study, we presented a method to design nanobioconjugates that constitute a new approach for novel QDs with long-lived PL emission. The presented QD-EuC-biotin nanobioconjugates extend the range of materials that can be applied in fluoroimmunoassays, imaging, flow cytometry, or microfluidic optical sensing systems. In such applications, the EuC-QDs can possibly improve the signal-to-noise ratio due to strongly reduced autofluorescence detection when compared to QDs with short PL decay times.