Vous êtes sur la page 1sur 11

DOI: 10.1002/cphc.

200500217

Frster Resonance Energy Transfer Investigations Using Quantum-Dot Fluorophores


Aaron R. Clapp,[a] Igor L. Medintz,[b] and Hedi Mattoussi*[a]
Frster resonance energy transfer (FRET), which involves the nonradiative transfer of excitation energy from an excited donor fluorophore to a proximal ground-state acceptor fluorophore, is a well-characterized photophysical tool. It is very sensitive to nanometer-scale changes in donoracceptor separation distance and their relative dipole orientations. It has found a wide range of applications in analytical chemistry, protein conformation studies, and biological assays. Luminescent semiconductor nanocrystals (quantum dots, QDs) are inorganic fluorophores with unique optical and spectroscopic properties that could enhance FRET as an analytical tool, due to broad excitation spectra and tunable narrow and symmetric photoemission. Recently, there have been several FRET investigations using luminescent QDs that focused on addressing basic fundamental questions, as well as developing targeted applications with potential use in biology, including sensor design and protein conformation studies. Herein, we provide a critical review of those developments. We discuss some of the basic aspects of FRET applied to QDs as both donors and acceptors, and highlight some of the advantages offered (and limitations encountered) by QDs as energy donors and acceptors compared to conventional dyes. We also review the recent developments made in using QD bioreceptor conjugates to design FRET-based assays.

1. Introduction: Frster Resonance Energy Transfer


Frster (or fluorescence) resonance energy transfer (FRET) involves the nonradiative transfer of excitation energy from an excited donor fluorophore, D, (after absorption of a higherenergy photon) to a ground-state acceptor fluorophore, A, brought in close proximity, which can radiatively emit a lowerenergy photon.[1, 2] FRET processes are driven by dipoledipole interactions and depend on the degree of spectral overlap between donor photoluminescence (PL) and acceptor absorption, and on the sixth power of the separation distance between the donor and acceptor pair, r.[2] Within the Frster formalism, the rate of nonradiative energy transfer is given by Equation (1), B QD I tD r 6    6 1 R 0 r tD The spectral overlap integral, I, defined as in Equation (3), I Z J l d l Z PLDcorr l l4 eA ldl 3

is a quantitative measure of the donoracceptor spectral overlap over all wavelengths l, where PLDcorr and eA represent the donor emission (normalized dimensionless spectrum) and acceptor absorption extinction coefficient spectrum, respectively. The above expressions imply that a quantum-yield measurement is necessary for deriving an accurate estimate of the Frster distance. The FRET efficiency, E, defined as in Equation (4): E kDA R6 0 1 6 R0 r 6 kDA tD 4

kDA

where tD is the excited-state radiative lifetime of the donor and R0 is the Frster separation distance corresponding to a rate of FRET equaling the rate of radiative decay (kDA = 1/tD). R0 is a function of the refractive index of the medium, nD ; Avogadro number, NA ; the donor PL quantum yield, QD ; the overlap integral, I ; and a parameter, kp, that depends on the relative orientation of the donor and acceptor dipoles; it is expressed as in Equation (2): 9000ln 10k2 p QD I NA 128p5 n4 D !1=6 2

accounts for the fraction of excitons transferred from donor to acceptor nonradiatively. This can be experimentally measured by monitoring changes in the donor or/and acceptor fluorescence intensities, or changes in the fluorescent lifetimes of the
[a] Dr. A. R. Clapp, Dr. H. Mattoussi US Naval Research Laboratory Optical Sciences Division, Code 5611, Washington, DC 20375 (USA) Fax: (+ 1) 202-404-8114 E-mail: hedi.mattoussi@ccs.nrl.navy.mil [b] Dr. I. L. Medintz US Naval Research Laboratory Center for Bio/Molecular Science and Engineering Code 6900, Washington, DC 20375 (USA)

R0

ChemPhysChem 2006, 7, 47 57

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

47

H. Mattoussi et al. fluorophores. During FRET, energy transfer manifests as a decrease in donor fluorescence intensity and shortening of its exciton lifetime; this may in turn be coupled with enhancement of the acceptor fluorescence and increase of its exciton lifetime. For additional details on the comprehensive derivation of energy transfer formalism, as described in the Frster theory, and a description of the methods for monitoring FRET, the interested reader is referred to a few specific references.[1, 2] In a configuration where a single donor interacts with multiple acceptors brought in close proximity, we have extended/modified the above analysis in the case where all acceptors are equally separated from a central donor, and derived an expression of the FRET efficiency, Equation (5): E n; r nR6 0 6 nR6 0 r 5 cal applications (e.g., fluoroimmuno- and FRET-based assays). Unique advantages offered by QD fluorophores include broad tunable absorption spectra with high molar extinction coefficients ( % 10100 that of organic dyes), narrow symmetric PL spectra (full width at half maximum % 2540 nm) spanning the UV to near-IR with high quantum yields, exceptional resistance to photo- and chemical degradation, high photobleaching thresholds and large achievable energy separation between excitation and emission lines (i.e. large effective Stokes shifts).[1014] Two properties are particularly appealing for developing FRET-based investigations: 1) the ability to tune the fluorescent emission (for example, as a function of core size for binary CdSe nanocrystals),[1012] which can allow better control of the spectral overlap with a particular acceptor,[3, 4] and 2) the ability to excite mixed QD populations at a single wavelength far removed (> 100 nm) from their respective emissions which could allow extensive multiplexing.[1012] Herein, we review the latest developments geared towards using QDs to design and implement FRET-based investigations, with some emphasis on those having biological applications. In particular, we address the advantages offered and limitations encountered by QDs as energy donors and acceptors. In applying the Frster formalism to analyze the nonradiative transfer of excitation energy in systems involving QD fluorophores (e.g. from an excited QD to a proximal ground-state dye acceptor), we approximate the QD excited state (or exciton) as an oscillating (point) dipole. This approximation is derived from the combination of two factors: 1) strong overlap of the electron and hole carrier wavefunctions in the nanocrystal (dot size is smaller than the bulk Bohr exciton size), and 2) the spatial extent of the carrier wavefunctions as well as the separation between electron and holes carriers inside the nanocrystals are much smaller than the wavelength of light. Furthermore, the treatment of nonradiative energy transfer within the Frster formalism (dipoledipole interactions) has provided excellent description of most experimental data collected from systems using QD fluorophores alone or in conjunction with conventional dyes.[3, 4, 1517]

where n is the number of acceptors surrounding a single donor.[3, 4] FRET is a powerful photophysical reporting technique. It has been extensively used in a variety of in vivo and in vitro biological investigations, including the monitoring of DNA hybridization and sequencing, protein conformation studies, diffusion dynamics, and the monitoring of intracellular receptorligand binding and cellular membrane dynamics.[2, 58] The power of this technique derives from its high intrinsic sensitivity to small changes (0.510 nm) in the separation distance and orientation between the donor and acceptor dipoles, which has led to the characterization of FRET in several reports as a spectroscopic ruler.[2]

2. Luminescent Quantum Dots versus Organic Fluorophores: Background


Most of the biological investigations using FRET for signal transduction have used two types of fluorophores: molecular (or organic) dyes and genetically encoded fluorescent proteins.[6] Emissive fluorophores can act as donors and acceptors, while dark quenching dyes can function as acceptors only. However, the use of these types of fluorophores has several limitations, including low resistance to chemical and photodegradation and low photobleaching thresholds. Their broad absorption/emission spectra limit the number of possible pairing choices, and thus multiplexing capabilities. Furthermore, variation of the absorption and/or emission spectra requires the use of distinct molecular labels with the associated synthetic and conjugation constraints. Fluorescent proteins have the advantage of being genetically encoded and allow, for example, easier intracellular loading, manipulation, and expression of fluorescence-based sensing and imaging in live cells, but tend to have low quantum yields and broad absorption/emission spectra.[9] Proteins containing fluorescent amino acids (tryptophan, tyrosine, and phenylalanine) have also been used as FRET donors.[2] Quantum-dot (QD) fluorophores have unique optical and spectroscopic properties that offer a compelling alternative to traditional fluorophores in several fluorescence-based biologi-

3. QD Fluorophores and FRET Investigations


3.1. QDs as FRET Donors The first demonstration that CdSe QDs (core only), as other organic fluorophores, can engage in efficient nonradiative energy transfer was reported by the Bawendi group.[15] The authors prepared thin films made of mixed close-packed QDs, grown by using high-temperature solution chemistry,[10] of two different diameters: % 38.5 (PL maximum at % 555 nm) functioning as exciton donors and larger QDs ( % 62 diameter, with PL maximum at % 620 nm) serving as energy acceptors. Solutions containing populations of these nanocrystals were prepared and used to spin-cast films of close-packed interpenetrating networks of donoracceptor nanocrystals. Steadystate fluorescence data (collected at room temperature and at 10 K) showed that there was a clear decrease in the PL from the 38.5- QDs along with an increase in acceptor (62- QDs)
ChemPhysChem 2006, 7, 47 57

48

www.chemphyschem.org

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Quantum-Dot Fluorophores contribution when exciting to the blue edge of the first absorption maximum of the smaller-size nanocrystals (Figure 1). Time-resolved fluorescence experiments corroborated the steady-state results where a decrease in the PL lifetime of the

Figure 1. Absorption and emission spectra of 38.5- ( % 555-nm emission peak) and 62- ( % 620-nm emission peak) CdSe QD close-packed solids at a) RT and at b) 10 K. PL spectra of the mixed system (solid lines) dispersed in solution at c) RT and d) 10 K and close-packed in the solid films at e) RT and f) 10 K. The system was made of 18 % of 62- dots mixed with 82 % of 38.5 nanocrystals; an excitation of 2.762 eV ( % 448 nm) was used. Dotted lines are the relative quantum yields for 38.5- dots in a pure film and for 62- dots in the mixed film when excited to the red (2.143 eV, % 578 nm) of the 38.5- dot absorption edge at e) RT and f) 10 K. Reproduced from ref. [15] with permission from the American Physical Society.

Figure 2. a) PL decays from a close-packed film of a single population of CdSeZnS nanocrystals with an average core CdSe radius of 12.4 and a ZnS shell of 9 at the energies specified in the inset. The inset shows steady-state PL spectra from the close-packed film (solid) and original dilute solution of the same QDs (dashed). b) Dynamic redshift of the peak emission; the inset shows the PL spectra at the specified times. Reproduced from ref. [17] with permission from the American Physical Society.

donor and an increase in the acceptor PL lifetime were measured; data were successfully analyzed within the Frster formalism.[15] In subsequent studies, the authors showed that in a population (with a finite size distribution) of close-packed CdSe QDs, there is shift of the PL maximum wavelength compared to the wavelength of their dilute solution counterparts, which was attributed to FRET from small- to larger-size nanocrystals. They further confirmed that the shift is more pronounced when the size distribution becomes broader.[16] In a more recent study, Crooker et al. investigated FRET dynamics in close-packed films of mixed-size and gradient (layered) assemblies of CdSe (core only) and CdSeZnS coreshell QDs.[17] In addition to measuring a red shift in the PL from close-packed films of core and coreshell QDs compared to the PL of their solution counterparts (as reported in ref. [16]), the authors employed time-resolved and spectrally-resolved fluorescence to probe the exciton decay acquired at specific energies (narrow wavelength windows) and showed that the PL decay was not constant along the emission energy spectrum. For a close-packed population of nanocrystals (average core radius = 12 and shell = 9 ), a short (1.9 ns) initial PL decay was measured at high energy (smaller-size nanocrystals), and it progressively increased with decreasing energy (or increasing QD size) to % 22 ns at the lowest energy probed (see Figure 2).[17] In contrast, a single average lifetime of % 24 ns was measured for a dilute solution of the same QD population.
ChemPhysChem 2006, 7, 47 57

They then fabricated close-packed films made of a mixture of three disparate-size TOP/TOPO-capped CdSe populations (radius = 17, 18, and 21 ) and compared the PL dynamics in such samples to that of their single-population film counter-

Figure 3. a) Steady-state PL spectra from close-packed solid films of 17- (~), 18- (*) and 21- (*) CdSe colloidal QDs, along with the spectrum from a film made of an equal mixture of the three populations of QDs (&). b) The corresponding PL lifetimes versus energy for each sample shown in (a). Reproduced from ref. [17] with permission from the American Physical Society.

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemphyschem.org

49

H. Mattoussi et al. parts (PL spectra are shown in Figure 3 a). They found that the emission from the close-packed solid mixture collapses into a single spectrum strongly weighted towards low energies or large sizes, while in a mixed dilute solution the PL spectrum is a composite sum of the three PL spectra. Furthermore, efficient FRET is seen in the three close-packed films of each population, through a pronounced decrease in PL lifetime from low to high energies, but the blue decay (700 ps) was % three times faster than in the CdSeZnS dots discussed above, a feature attributed to the smaller donoracceptor separation distance in core-only QDs. When the three-size QDs are combined in a mixed film, they measured short PL lifetimes (700 800 ps) throughout the high-energy range (covering the two smaller sizes), whereas at low energies longer lifetimes were measured, roughly mimicking the behavior of the film with the largest size QDs. These studies demonstrated that QDs could potentially function both as FRET donors and acceptors. In addition, for QD population of a particular size, the rate of FRET is due primarily to the availability of nearby viable acceptor dots. Center-tocenter separation distances for these QD films are only slightly larger than R0 values, which confirms that significant FRET would occur in these systems. Following the early demonstrations of water-soluble QDs as potentially useful fluorophores for developing biological assays,[1820] interest in developing assays based on QDs and FRET naturally followed. Willard et al. reported the first biologically relevant FRET investigation using QDs as energy donors.[21] Water-soluble CdSeZnS QDs were first conjugated to biotinylated bovine serum albumin (bBSA, with % 11 bBSA per QD) via a thiol linkage, although free bBSA was still present in solution. In a separate reaction, StreptAvidin (SAv) was covalently labeled with tetramethylrhodamine (TMR) and used as acceptor. The two subunits were allowed to react in phosphate buffered saline (PBS) to form QDbBSASAvTMR complexes via avidinbiotin interactions. Emission spectra were measured for each subunit separately and for the complex as a function of SAvTMR concentration. They reported a decrease in the QD PL and an increase in the TMR PL as the molar ratio of SAvTMR to QD increased from 0 to 11 (presumably about the number of bound bBSA per QD), and attributed their results to FRET between QDs and TMR on the SAv. No time-resolved fluorescence experiments were carried out and no estimates for the separation distance or for R0 were provided for this QD dye pair. Approximate estimates of these values using available experimental data indicate that R0 and r are % 54 and 100 (bBSA minor axis % 60 , SAv minor axis % 48 ), respectively, suggesting that low FRET efficiencies would be expected with this QDproteindye configuration. In a context similar to the one used by Kagan et al. but adapted to a biological setting, a recent study reported that by using green-emitting CdTe QDs conjugated to anti-BSA antibody and red-emitting QDs conjugated to BSA, assembled together via receptorligand interactions, resulted in quenching of the green QDs and an enhancement in the luminescence of the red QDs. Results were again attributed to FRET between QDs of different sizes brought together by specific interactions.[22] This configuration produces a rather large center-tocenter separation distance between donor and acceptor, and in the absence of time-resolved fluorescence data interpretation based only on FRET may not be fully warranted. Our group took a more systematic approach to understanding FRET between CdSeZnS QDs and proximal dyes. For this, QD donors capped with dihydrolipoic acid (DHLA) were conjugated to engineered proteins containing site-specifically labeled dye acceptors and appended with either a C-terminal oligohistidine (His), or a basic leucine zipper (zb) attachment;[3, 4, 20, 23, 24] Figure 4 A schematically depicts the bioconjugate structure. In such configuration, each QD is surrounded by a fixed number of proteins (e.g., % 5, 10 or 15 maltosebinding proteins, MBPs); the proteins adhere onto the nanocrystal surface through either electrostatic self-assembly (for MBPzb) or via metal-affinity coordination between the QD surface and C-terminal histidine (for MBP-His).[3, 4, 20, 23, 24] The total number of MBP bound to a QD remained fixed while we varied the number of dye-labeled proteins. More importantly, such configuration maintains the average distance between the QD center and the proximal dyes in each conjugate fixed. The proteins have the same anchoring point for attachment to the QD, which ensured consistent protein orientation and donoracceptor distances. These studies utilized different size/ color QDs whose emission maxima were varied to tune/optimize the degree of spectral overlap with the acceptor absorption (see Figure 4 B). Steady-state fluorescence measurements showed that as the fraction of dye-labeled proteins per QD increased the QD PL systematically decreased.[3, 4, 24] In the case where Cy3 (an emitting dye) was the acceptor a concomitant enhancement of Cy3 PL was measured (Figure 4 CD).[3] Similarly, time-resolved fluorescence experiments on these self-assembled QDproteindye conjugates showed a progressive decrease of the QD exciton lifetime with increasing dye-to-QD ratio (Figure 4 E).[3] These two sets of observations confirm that efficient nonradiative exciton transfer between QD donors and dye acceptors is realized by using dye-labeled proteins self-assembled on QD surfaces, and that the overall behavior is consistent with the Frster formalism. We further analyzed the FRET data collected from QD PL loss (or dye PL gain) with increasing Cy3-to-QD ratio in each conjugate using Equation (5) to extract estimates of the separation distance, r, at each ratio n.[3] Moreover, compilation of the values at each ratio (derived for each QDCy3 pair) provided an average distance with a statistical error. We found that the average values for r were generally consistent with those anticipated by using the radius of the QD inorganic core, the dimension of the proteins used, and assuming a close approach between QD and proteindye in a self-assembly process. In a recent study, Grecco et al. used commercially available CdSeZnS QDs emitting at 585 nm coated with a thick passivating and functionalized polymer layer and conjugated to SAv as FRET donors with proximal AlexaFluor 594-biocytin; specific interactions between biocytin and SAv brought the acceptor in close proximity to the QD center.[25] In particular, they measured FRET efficiencies that increased (though linearly)
ChemPhysChem 2006, 7, 47 57

50

www.chemphyschem.org

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Quantum-Dot Fluorophores

Figure 4. A) Schematic representation of a QDproteindye assembly (not drawn to scale). The QD consists of a CdSeZnS coreshell capped with DHLA. The charged carboxyl on the DHLA keeps the colloid nanocrystals dispersed in solution at basic pH. Also depicted are representative proteins (MBP, in this case) that are immobilized on the QD via metal-affinity coordination. Site-specific dye labeling of the MBP allows a quasi-symmetrical arrangement of Cy3, at a fixed distance, r, from the QD center. B) Normalized absorption spectra (eA) of Cy3 dye and photoemission spectra of the three CdSeZnS coreshell QD solutions. The inset shows plots of the resulting overlap functions J(l) [Eq. (3)] and highlights the effects of tuning the QD emission on the degree of spectral overlap with Cy3 dye. C) Evolution of the photoluminescence spectra derived from titrating 510-nm emitting QDs with an increasing ratio, n, in QDMBPCy3 assemblies. Spectra have been deconvoluted and corrected for direct excitation contribution to acceptor emission; this allows discrimination in QD donor loss and Cy3 acceptor gain. D) Integrated intensities for both QD and Cy3 versus ratio n. E) Series of false-color timewavelength intensity images showing the intensity of: free MBPCy3, top series; 510-nm QDs with 10 unlabeled MBP per QD, middle series; and 510 nm QDs with 3 MBP/7 MBPCy3 on the QD surface, bottom series; as recorded by a CCD camera at 2-ns intervals following an initial 90-ps laser pulse. Horizontal axis represents wavelength; dashed lines mark the maximal emission for the QD (at 510 nm) and the dye (at 570 nm). Adapted from ref. [3] with permission from the American Chemical Society.

with Alexa 594-to-QD ratio, similar to what we reported in ref. [3, 4]. However, reported efficiencies were much higher than what was anticipated from estimates of R0 and the relatively large separation distances involved. Furthermore, given the multivalency of the receptors on the QDs, accurate control over donoracceptor separation distance in their system could not be achieved. The authors attributed these observations to
ChemPhysChem 2006, 7, 47 57

possible aggregate formation and overestimation of the initial QD concentration.[25] 3.2. QDs as FRET Acceptors In principle, QDs should be excellent FRET acceptors due to their large absorption cross-section, which can be an order of

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemphyschem.org

51

H. Mattoussi et al. magnitude larger than that of organic dyes.[3, 13] Studies by Kagan and Crooker demonstrated that QDs can transfer energy nonradiatively between populations of different sizes,[1517] but their ability to accept excited-state energy from organic dyes or other donor fluorophores was unclear. In a biological context (using QD bioconjugates), there are few examples of QDs functioning as energy acceptors in the literature. Mamedova et al. described the use of 1:1 conjugates of BSA and 580-nm emitting CdTe QDs and reported substantial energy transfer from the native tryptophan (Trp) residues in BSA to attached QDs; the system was excited at 290 nm.[26] However, it is widely known that addition of surface-bound protein alone can significantly influence the PL of QDs in solution irrespective of any potential FRET interaction.[4, 20, 26, 27] In the absence of suitable control experiments, it is not clear whether the above observations could unequivocally be attributed to FRET or other phenomena such as protein-induced surface passivation of QD trapping states. We explored the use of QDs as FRET acceptors in our QD MBP system by labeling MBP with various organic dyes that emit to the blue edge of the QDs.[28] Excitation wavelengths that maximize dye excitation and emission also resulted in substantial direct excitation of the QD acceptors due to their broad absorption spectra. Using AlexaFluor 488 (AF488) and Cy3 dyes as donor fluorophores and 555-, 570-, and 590-nmemitting QDs as acceptors, we found no evidence of nonradiative energy transfer from dye to QDs. Steady-state and time-resolved experiments showed no apparent change in fluorescence behavior from either donor or acceptor, as compared with the fluorescence of the independent control samples. Similar to the above study,[26] we also attempted Trp-to-QD FRET by exciting QD-bound MBP at 280 nm (each MBP molecule contains % 120 Trp residues). Our results showed that there was no evidence of Trp quenching or significant QD PL enhancement when MBP was placed near the QD surface. As expected, the addition of MBP alone was found to be solely responsible for observed QD PL enhancement (due to surface passivation effects), which suggests that FRET was absent in this system even though donoracceptor separation distances were favorable (r % 1.3 R0).[28] Our results were attributed to the slow decay rate of the QDs combined with a strong direct excitation.[28] Lifetimes of soluble CdSeZnS QDs were reported to vary between 3 and 15 ns depending on the conditions used.[3, 15, 17, 28] To test whether relative donoracceptor lifetimes were important parameters for efficient FRET, we replaced the above organic fluorophore with a long-lifetime Ru-bpy isothiocyanate (ITC) metal chelate dye covalently attached to the MBP. Steady-state data were inconclusive, due to a pronounced overlap in donor and acceptor PL spectra and low Ru-bpy quantum yield. However, time-resolved fluorescence data showed a substantial decrease in Ru-bpy lifetime when bound to the QD surface, indicating FRET from Ru-bpy to the central QD (see Figure 5).[28] Although not fully conclusive, the results provide strong evidence that the lifetime of the acceptor relative to that of the donor plays a significant factor for generating efficient FRET for a given donoracceptor pair.

Figure 5. Fluorescence intensity plotted as a function of delay time following a 90-ps excitation pulse: A) MBPRu-bpyITC alone and B) MBPRu-bpy ITC-610QD conjugate. The insets are timewavelength intensity plots for each sample with the same false-color intensity scale. A fit of the data shown in (A) provides a long exciton lifetime of 420 ns, whereas in (B) two distinct contributions can be isolated: a relatively short lifetime (attributed to the QDs and a long lifetime attributed to the Ru-bpyITC. The short decay time ( % 7 ns) derived from the data by the fitting procedure is identical to the one for the MBPQD-only sample, whereas the longer decay time ( % 160 ns) is about half of the lifetime measured for MBPRu-bpyITC alone. Adapted from ref. [28] with permission from the American Chemical Society.

The above findings contrasted with results from two recent studies suggesting that luminescent QDs can be effective energy acceptors with polymer and quantum well (QW) donors. In the first study, Acherman et al. reported efficient nonradiative energy transfer to QDs in a three-layer heterostructure made of an InGaN QW donor and an adjacent layer made of close-packed CdSeZnS QDs.[29] The QD layer is deposited on top of the QW by using the LangmuirBlodgett technique, either directly or separated by a thin GaN cap layer. The system was excited with a 200-fs laser pulse at 266 nm (third harmonic of a Ti-sapphire laser). The authors first demonstrated that the QW photoemission has a quadratic dependence on the excitation power, while its lifetime decay was independent of that power, which indicates that the InGaN QW is characterized by free-carrier recombination not bound excitons, due to weak electronhole interactions (Figure 6 A). This implies that the energy-transfer rate of carriers from the QW to QD layer
ChemPhysChem 2006, 7, 47 57

52

www.chemphyschem.org

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Quantum-Dot Fluorophores with the radiative and nonradiative decay rates of the QW excitation energy. The authors further suggest that the system could be modified to allow for electrical excitation of the QW, followed by energy transfer and emission from the QD layer.[29] In the second report, the authors performed low-temperature (at 18 K) steady-state and time-resolved fluorescence experiments using thin films made of a blue fluorescing polymer matrix doped with QDs at low loading concentrations; with a QD PL centered at 560 nm and a broad polymer fluorescence with a composite peak between 420 and 500 nm, R056 for this pair. Samples were excited with a short (2 ps) laser pulse at 390 nm (the second harmonic of a mode-locked Ti-sapphire laser).[30] Time-resolved measurements, limited to time scales below 1 ns, indicated that as QD concentration increased a shortening of the polymer exciton lifetime along with an apparent lengthening of the QD excitation lifetime could be measured, which was attributed to an efficient FRET from polymer to QDs. Analysis of the efficiencies extracted from the time-resolved data, allowing for exciton diffusion contribution in the rate equations (since the entire polymer matrix is the donor), provided an experimental value for R0 of 80 , about twice the value estimated from the overlap integral (R056 ). This implies that higher than expected FRET efficiencies were measured for this system. The authors attributed this difference to an error in estimating the polymer molecular weight and density. Steady-state fluorescence showed a clear contribution from the QDs at the larger concentration, but the data could not be corrected for a direct excitation contribution, which makes an interpretation based only on FRET ambiguous. It is difficult to provide an accurate and objective comparison between the findings discussed in these studies with those described above in ref. [28]. The QW donor used in ref. [29] has different carrier characteristics than our organic dyes, and the energy-transfer rates predicted for that system are higher than those predicted for single donoracceptor pairs (Equation (3)). In ref. [30], the time-resolved data were collected at low temperature by using a shorter excitation pulse and analyzed within a short time window (t < 1 ns), whereas in ref. [28] data were collected at room temperature, by using a longer excitation pulse (90 ps) and analyzed over a longer time window (0.5 ns < t < 50 ns with 0.5 ns resolution). Additional studies are necessary to clarify the issues associated with QDs as energy acceptors with organic donors. Better control samples of dots only and dye or polymer only should be used in both steady-state and time-resolved experiments. Exploring the effects of donor lifetime on the occurrence of FRET and its efficiencies would also be extremely useful in providing better understanding of these systems.

Figure 6. Experimental observations of QW-to-nanocrystal(NC)-layer energy transfer. a) Normalized PL dynamics of the isolated QW (black solid line) at carrier density of neh = 3 012 cm2 in comparison with QW PL dynamics measured for the QW/QD structure at neh = 3 012 cm2 (grey dotted line) and at neh = 10 012 cm2 (grey dashed line). b) The difference between the initial PL decay rates measured for the isolated QW and the QW/QD structure (DG = GQW with NCGQW without NC) versus QW carrier density for samples based on the capped and the uncapped QW. c) Time-integrated nanocrystal PL intensity versus pump fluence for the nanocrystal monolayer assembled on a glass substrate and on top of a capped QW. Reproduced from ref. [29] with permission from Nature Publishing Group.

varies as 1/d4 (where d is the separation distance between the QW and QD layer centers). Time-resolved measurements showed a power-dependent shortening of the QW PL lifetime in the presence of the QD layer in comparison with the QW lifetime alone. Furthermore, changes in the PL decay rates for the two-layer structure compared to the rates of the QW alone (DG = GQW with NC(QD) layerGQW alone) varied linearly (in a semilogarithmic plot) with the electronhole carrier density (neh), but were consistently smaller in the presence of the cap GaN layer. Using a single layer of QDs deposited on a glass substrate as a control, they also showed that at high excitation power the three-layer structure provided a larger PL intensity than the control sample with saturation occurring for both sample and control. The difference in PL signal also increased with laser power before saturation (Figure 6 C). The authors attributed these observations to efficient nonradiative energy transfer between the QW and the top QD layer. Energy transfer is more efficient because of its inverse fourth-power dependence on the separation distance (E % 1/d4), as expected for a donor system with unbound (or very weakly bound) electronhole carriers. This results in a fast energy-transfer rate compared
ChemPhysChem 2006, 7, 47 57

4. Bio-Inspired FRET-Based Applications Using QD Donors


4.1. Use of FRET to Derive QDProtein-Conjugate Configuration A property of crucial importance when assembling QDbioreceptor conjugates is the final bioreceptor orientation within

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemphyschem.org

53

H. Mattoussi et al. the conjugate, since this has ramifications on the conjugate biological functionality; proteins that are oriented on the QD heterogeneously will manifest a mixed functional avidity and may not perform optimally in biological assays. With this in mind, we set out to derive the optimal configuration of our self-assembled conjugates using FRET between a central QD donor and a dye acceptor whose position in the protein was selectively controlled.[31] His-terminated MBP mutants expressing six individual and unique cysteine residues spread across the two protein lobes were site-specifically labeled with rhodamine red (RR) and immobilized on the QDs (Figure 7 A). QD MBPRR conjugates with different RR-to-QD ratios (as described above) were used and average distances (ri) from the QD center to each RR location on the protein were derived from FRET data. Applying the concepts developed for a global positioning system with its spherical coordinate symmetry, but adapted to our nanoscacle QDprotein conjugate (nanoscale GPS), a least squares minimization of the deviations between the experimental distances (ri, derived from the FRET data) and those derived from the protein PDB crystallographic structures was carried out for all six dye locations. This model assumed a protein that is free to translate on the QD surface and free to rotate around the conjugate polar axis originating at the QD center. We were able to accurately determine the optimal configuration of the self-assembled QDMBP conjugates (Figure 7 B). A critical result from this analysis was that MBP self-assembles onto QD surfaces with a preferred homogeneous conformation where the C-terminus His tail in each MBP comes in close contact with (coordinates to) the QD, while the binding pocket faces away from that surface, thus preserving protein function in the conjugate.[31] This approach of using self-assembly combined with FRET to derive protein conformation in a functional bioconjugate could be applied to a wide range of QDprotein conjugates. The derived information could potentially be used to infer protein structure as it self-assembles on the nanocrystals and its ultimate functionality. 4.2. Photochromic Switching Based on QD FRET Using a QD donor and photochromic dye acceptor pair, we tested the ability of such system to realize modulation of FRET rates induced by dye photochromicity (photochromic FRET), and use that to control QD emission.[32] To do this, we labeled MBP with a sulfo-N-hydroxysuccinimide activated photochromic BIPS molecule (1,3-dihydro-1-(2-carboxyethyl)-3,3-dimethyl-6-nitrospiro[2H-1-benzopyran-2,2-(2 H)-indoline]) at a dye-toMBP ratio of % 5. BIPS-labeled MBPs were self-assembled on the QDs and the ability of BIPS to modulate QD photoluminescence was tested by switching the BIPS from the colorless spiropyran (SP nonabsorbing) to the colored merocyanine (MC energy absorbing). FRET efficiency for this system was measured following irradiation with white light (> 500 nm) and UV light ( % 365 nm), which respectively generate the SP and MC forms of the dye. A FRET efficiency of % 60 % was measured in the presence of the MC form compared to % 10 % for the SP form. This system has also shown the benefit of arraying multiple acceptors around a single donor, where starting with a donoracceptor pair having a modest overlap (R038.5 ) and immobilizing 20 MBPs ( % 100 dyes) on each QD produced large experimental FRET efficiencies.[32] 4.3. QDProtein-Sensing Assemblies Based on FRET Building on our understanding of optimal FRET conditions for QDproteindye conjugates, we constructed a prototype sensor for the specific detection of the nutrient sugar maltose in solution using a QDMBP conjugate. This sensing assembly used QDs as both energy donors and structural scaffolds to array multiple protein receptors (see Figure 8).[4] The equilibrium dissociation constant for maltose, KD, derived from saturation data agreed well with literature values for the solutionChemPhysChem 2006, 7, 47 57

Figure 7. A) 3-D representation of an MBP protein with the labeled residues highlighted and numbered. The location of the His tail is also shown. B) Refined arrangement/orientation of MBPHis as it self-assembles on the surface of a QD. Side view presenting the structure of the final refined MBP orientation with all six dye locations highlighted in red. The refined distances from each of these dyes to the center of the spheroid are shown in yellow. Adapted from ref. [31] with permission from the Proceedings of the National Academy of Sciences.

54

www.chemphyschem.org

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Quantum-Dot Fluorophores nal 12-His sequence was allowed to prebind an analog of TNT (TNB which is conjugated to the dark quenching dye BHQ10) and then attached to the QD surface (similar to the scheme shown in Figure 8 A). Again, due to excellent spectral overlap and close proximity of the QD donor to the BHQ10-quenching acceptor dye, a progressive and BHQ10-dependent quenching of QD fluorescence is measured. As TNT is added to the assay solution, it competes for binding to the antibody fragment and the QD fluorescence increases as TNBBHQ10 is displaced in a concentration-dependent manner. Antibody-fragment specificity for explosive analogs of TNT was retained in this QD sensing configuration.[33] Both of these studies demonstrate that QDs not only function as efficient FRET donors, but also as scaffolds for assembling the sensor elements. 4.4. QD FRET To Probe DNA Replication and Telomerization Replication of DNA is an important natural process that occurs in live cells before they divide. It was shown that replication on solid surfaces could be employed to amplify DNA detection.[34] QDs conjugated to oligonucleotides have been used as FRET donors to monitor DNA replication and telomerization processes in solution.[35] CdSeZnS QDs were attached to a thiol-terminated DNA primer (sequence: 5-HS-(CH2)6CCCCCACGTTGTAAAACGACGGCCAGT-3), then incubated with the complementary sequence M13 fDNA to allow for hybridization. Addition of polymerase (to initiate replication) mixed with dNTPs and dUTP labeled with Texas Red (TRdUTP) permitted to follow the dynamics of DNA replication with time, by monitoring changes in the QD and dye emissions. Replication of the DNA progressively brings TRdUTP complexes specifically in close proximity to the QD center, resulting in time-dependent FRET between the QD donor and proximal TR acceptors. In particular, FRET data show that the replication process starts fast (nearly linearly with time) and saturates after 1 hour (Figure 9). The authors further applied FRET-induced quenching of QD emission to detect the dynamics of DNA telomerization, by incubating the QDDNA primer conjugates with dNTP mixtures that include TR-labeled dUTP and telomerase.[35] FRET efficiencies were relatively weak for this system due the rather large separation distance between QD and acceptors involved; further, the nature of this system does not allow accurate control over separation distance. 4.5. QDs in Photodynamic Medical Therapy phase-behavior of wild-type MBP.[4] Time-resolved fluorescence data collected for this system showed a decrease in the QD lifetime when the dye-labeled analog was bound to MBP, and recovery of the QD lifetime when maltose was added to the system. The nanosensor also shows high specificity by responding only to sugars having the MBP-recognized a-1,4-glucosidic linkage, which proves that QD-bound proteins maintain their intrinsic binding properties.[4] A second sensing assembly based on FRET and QD energy donors targeted the explosive TNT.[33] For this, an antibody fragment that recognizes TNT (TNB2-45) expressing a C-termiChemPhysChem 2006, 7, 47 57

Figure 8. A) Schematic description of functions and properties of a solutionphase QDMBPHis-sensing assembly targeting maltose. Each QD is surrounded by an average of % 10 MBP moieties. b-CD (beta-cyclodextrin) analog attached to a QSY-9 quenching dye is prebound in the MBP binding pocket. Formation of QDMBPb-CDQSY-9 complex brings the dye in close proximity to the QD center and results in substantial FRET quenching of QD emission due to nonradiative energy transfer from QD to QSY-9; the sensor is in the initial off state. B) Added maltose displaces b-CDQSY-9 away from the QD donor, which results in a concentration-dependent increase in QD emission. The saturation curve (derived from the titration of QDMBP conjugate (preassembled with 1 mm b-CDQSY-9) with increasing concentrations of maltose is shown. Adapted from ref. [4] with permission from Nature Publishing Group.

Photodynamic therapy is a process in which an excited photosensitizing agent transfers its excitation energy to a nearby oxygen molecule to form a reactive singlet oxygen (1O2). This highly reactive species causes cytotoxic reactions in cells or tissues, which has made the process a useful therapeutic tool to treat cancerous tissue and cells. The technique is highly selective because only tissues that are simultaneously exposed to the photosensitizer and photoexcitation in the presence of oxygen are affected. In a preliminary study, QDs were linked to a known silicon photosensitizer (Pc4) through alkyl amine groups. Photosensitization of the Pc4 alone is limited to the

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemphyschem.org

55

H. Mattoussi et al. based FRET using lipid membranes and surface-immobilized assemblies have potential use for monitoring protein interactions and dynamics near cellular membranes,[41] or constructing surface-tethered sensing assemblies using QDs as a scaffolds and exciton donors.[42]

5. Summary and Conclusions


Figure 9. A) changes in emission spectra upon the time-dependent DNA replication. Before addition of TRdUTP, orange; after 1, 30, and 60 min of replication, black, purple, and blue, respectively. B) Time-dependent intensity of the QD and TR during replication: decrease of QD emission (&), increase of dye contribution (*). Adapted from ref. [35] with permission from the American Chemical Society.

window between 550 and 630 nm (absorption peak of the Pc4). Taking advantage of the broad excitation spectra of QDs, the authors showed that one can achieve photosensitization of the Pc4 via FRET between the QD center and proximal Pc4 groups; the system was excited at 488 nm, far from the absorption peak of Pc4.[36] Furthermore, due to the short donor acceptor separation distance high FRET efficiencies and thus high degrees of photosensitization were achieved. The same study has also shown that QDs alone can generate 1O2 in the absence of Pc4, detected by measuring and analyzing emission at 1270 nm; yields from this process were much lower that those generated with Pc4, however. 4.6. Additional Uses of QD-Based FRET There have been other FRET-based investigations (bio-inspired and otherwise) using QDs as energy donors where authors have, for example, arrayed multiple acceptors around a single nanocrystal or selected the excitation line over a broad range of wavelengths.[3743] They include the use of QDs as exciton donors and scaffolds for DNA molecular beacons, where changes in FRET rates to dyes attached at the beacon end were assessed, and use of FRET modulation to demonstrate an inhibition assay of binding between SAvQDs and proximal biotinylated Au nanoparticles in the presence of molecular inhibitors.[37, 38] There is also growing interest in realizing FRET at the single-molecule level, where microcopy (optical and scanning probe) techniques combined with spectroscopy are used to image individual QDs interacting with proximal dyes or Au nanoparticle quenchers. In one study, AFM tips were functionalized with QDs, and their potential as highly localized FRET imaging systems was demonstrated by bringing a dye acceptor close to the tip and measuring distinctive FRET signals.[39] In another study, the authors used double-stranded DNA as a rigid spacer to tune the distance between a QD and proximal Au nanoparticles and reported distance-dependent quenching of QD emission.[40] Energy transfer between lipid membranes, or between QDs and dye-labeled proteins assembled on a surface have also been investigated.[41, 42] Demonstration of QD-

Herein, we reviewed the developments and progress made in the past few years using luminescent QDs as energy donors and acceptors in FRET investigations, both in purely spectroscopic and biological contexts. As energy donors, QDs offer several unique advantages compared to conventional dyes. These include the ability to tune the degree of spectral overlap with a given dye acceptor by, for example, varying the nanoparticle size for binary materials. In addition, due to their large size (comparable to that of a protein) QDs provide a unique configuration where multiple dye-labeled bioreceptors can be arrayed around a single nanoparticle, which enhances the effective FRET cross-section, even for a modest overlap (i.e., small Frster distance). Both of these properties increase the measured FRET rates and improve the signal-to-noise ratio. Further, the possibility of arraying several receptors with different properties/functionalities on QD donors provides the potential for multiplexing. These results indicate that FRET applications using QDs could see significant use both in biological and nonbiological contexts. Accurate derivation of protein conformation and sensor design represent only two of a long list of potential applications. QD-based FRET nanosensors will be particularly appealing for intracellular sensing, where their high photobleaching thresholds and substantial reduction in direct excitation of dye and fluorescent protein acceptors could permit the monitoring of intracellular processes over longer periods of time. FRET-based uses of QDs will potentially find applications in the monitoring of a wide array of intra- and extracellular indicators; success will still depend on the ability to design easy and reproducible means of delivering compact QDbioreceptor conjugates specifically to subcellular regions and cell membranes. This could also provide a very useful tool to investigate protein dynamics and interactions in vivo over extended periods of time. Further development of FRET assays using QDs and QD conjugates tethered to surfaces could open up the possibility of designing and implementing regenerable sensing assemblies and devices. However, there remain limitations to using these inorganic fluorophores in FRET-based investigations, most notably in biological applications. Due to their large and broad excitation spectra and long exciton lifetimes, they may not be useful as acceptor fluorophores. Their relatively large size and the need
ChemPhysChem 2006, 7, 47 57

56

www.chemphyschem.org

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Quantum-Dot Fluorophores for surface functionalization can increase the overall dimensions of the bioconjugates. Despite the substantial progress made in the last several years to prepare water-compatible and functional well-dispersed QDs, problems remain including aggregation, limited pH stability, and large QDbioreceptor conjugate sizes, that still need to be solved. Unfortunately, a few publications reporting FRET with luminescent QDs lacked necessary control experiments to prove unequivocally that energy transfer was the only means of interaction and not other uncontrolled processes. As progress is made in preparing additional QD materials and designing additional QD bioconjugates with good control over conjugate architecture and functionality, better understanding of energy transfer applied to luminescent QDs and further refinement of FRET-based QD sensing assemblies that are specific will be developed.
[16] a) C. R. Kagan, C. B. Murray, M. G. Bawendi, Phys. Rev. B 1996, 54, 8633 8643; b) C. R. Kagan, Ph. D. Dissertation, Massachusetts Institute of Technology, 1997. [17] S. A. Crooker, J. A. Hollingsworth, S. Tretiak, V. I. Klimov, Phys Rev. Lett. 2002, 89, 186802. [18] M. Bruchez, M. Moronne, P. Gin, S. Weiss, A. P. Alivisatos, Science 1998, 281, 2013 2015. [19] W. C. W. Chan, S. Nie, Science 1998, 281, 2016 2018. [20] H. Mattoussi, J. M. Mauro, E. R. Goldman, G. P. Anderson, V. C. Sundar, F. V. Mikulec, M. G. Bawendi, J. Am. Chem. Soc. 2000, 122, 12142 12150. [21] D. M. Willard, L. L. Carillo, J. Jung, A. Van Orden, Nano Lett. 2001, 1, 469 474. [22] S. Wang, N. Mamedova, N. A. Kotov, W. Chen, J. Studer, Nano Lett. 2002, 2, 817 822. [23] J. F. Hainfeld, W. Liu, C. M. R. Halsey, P. Freimuth, R. D. Powell, J. Struct. Biol. 1999, 127, 185 198. [24] P. T. Tran, E. R. Goldman, G. P. Anderson, J. M. Mauro, H. Mattoussi, Phys. Stat. Sol. B 2002, 229, 427 432. [25] H. E. Grecco, K. A. Lidke, R. Heintzmann, D. S. Lidke, C. Spagnuolo, O. E. Martinez, E. A. Jares-Erijman, T. M. Jovin, Microscopy Res. Technique 2004, 65, 169 179. [26] N. M. Mamedova, N. A. Kotov, A. L. Rogach, J. Studer, Nano Lett. 2001, 1, 281 286. [27] K. Hanaki, A. Momo, T. Oku, A. Komoto, S. Maenosono, Y. Yamaguchi, K. Yamamoto, Biochem. Biophys. Res. Comm. 2003, 302, 496 501. [28] A. R. Clapp, I. L. Medintz, B. R. Fisher, G. P. Anderson, H. Mattoussi, J. Am. Chem. Soc. 2005, 127, 1242 1250. [29] M. Achermann, M. A. Petruska, S. Kos, D. L. Smith, D. D. Koleske, V. I. Klimov, Nature 2004, 429, 642 646. [30] M. Anni, L. Manna, R. Cingolani, D. Valerini, A. Creti, M. Lomascolo, App. Phys. Lett. 2004, 85, 4169 4171. [31] I. L. Medintz, J. H. Konnert, A. R. Clapp, I. Standish, M. E. Twigg, H. Mattoussi, J. M. Mauro, J. R. Deschamps, Proc. Nat. Acad. Sci. USA, 2004, 101, 9612 9617. [32] I. L. Medintz, S. A. Trammell, H. Mattoussi, J. M. Mauro, J. Am. Chem. Soc. 2004, 126, 30 31. [33] E. R Goldman, I. L. Medintz, J. L. Whitley, A. Hayhurst, A. R. Clapp, H. T. Uyeda, J. R. Deschamps, M. E. Lassman, H. Mattoussi, J. Am. Chem. Soc. 2005, 127, 6744 6751. [34] F. Patolsky, Y. Weizmann, I. Willner, J. Am. Chem. Soc. 2002, 124, 770 771. [35] F. Patolsky, R. Gill, Y. Weizmann, T. Mokari, U. Banin, I. Willner, J. Am. Chem. Soc. 2003, 125, 13 918 13 919. [36] A. C. Samia, X. Chen, C. Burda, J. Am. Chem. Soc. 2003, 125, 15 736 15 737. [37] J. H. Kim, D. Morikis, M. Ozkan, Sensors Actuators B 2004, 102, 315 319. [38] E. Oh, M. Y. Hong, D. Lee, S. H. Nam, H. C. Yoon, H. S. Kim, J. Am. Chem. Soc. 2005, 127, 3270 3271. [39] Y. Ebenstein, T. Mokari, U. Banin, J. Phys. Chem. B 2004, 108, 93 99. [40] Z. Gueroui and A. Libchaber, Phys Rev. Lett. 2004, 93, 166 108. [41] J. A. Kloepfer, N. Cohen, J. I. Nadeau, J. Phys. Chem. B 2004, 108, 17 042 17 049. [42] K. E. Sapsford, I. L. Medintz, J. P. Golden, J.R Deschamps, H. T. Uyeda, H. Mattoussi, Langmuir 2004, 20, 7720 7728. [43] F. Mller, S. Gtzinger, N. Gaponik, H. Weller, J. Mlynek, O. Benson, J. Phys. Chem. B 2004, 108, 14 527 14 534. Received: April 18, 2005 Published online on December 21, 2005

Acknowledgements
The authors acknowledge NRL and ONR for support. A.R.C. is supported by a National Research Council Fellowship through the Naval Research Laboratory. H.M. acknowledges Drs. A. Ervin and L. Chrisey at the Office of Naval Research (ONR) and A. Krishnan at DARPA for research support. Keywords: bioconjugates fluorescence FRET (fluorescence resonance energy transfer) nanocrystals quantum dots
[1] T. Frster, Discuss. Faraday Soc. 1959, 27, 7 17. [2] J. R. Lakowicz, Principles of Fluorescence Spectroscopy, 2nd Ed. , Kluwer Academic, New York, 1999. [3] A. R. Clapp, I. L. Medintz, J. M. Mauro, B. R. Fisher, M. G. Bawendi, H. Mattoussi, J. Am. Chem. Soc. 2004, 126, 301 310. [4] I. L. Medintz, A. R. Clapp, H. Mattoussi, E. R. Goldman, J. M. Mauro, Nat. Mater. 2003, 2, 630 638. [5] E. A. Jares-Erijman, T. M. Jovin, Nat. Biotech. 2003, 11, 1387 1395. [6] A. Miyawaki, Developmental Cell 2003, 4, 295 305. [7] P. Wu, L. Brand, Anal. Biochem. 1994, 218, 1 13. [8] V. V. Didenko, Biotechniques 2001, 31, 1106 1121. [9] M. Fehr, W. B. Frommer, S. Lalonde, Proc. Natl. Acad. Sci. USA, 2002, 99, 9846 9851. [10] C. B. Murray, D. J. Norris, M. G. Bawendi, J. Am. Chem. Soc. 1993, 115, 8706 8715. [11] M. A. Hines, P. Guyot-Sionnest, J. Phys. Chem. B 1996, 100, 468 471. [12] B. O. Dabbousi, J. Rodriguez Viejo, F. V. Mikulec, J. R. Heine, H. Mattoussi, R. Ober, K. F. Jensen, M. G. Bawendi, J. Phys. Chem. B 1997, 101, 9463 9475. [13] C. A. Leatherdale, W. K. Woo, F. V. Mikulec, M. G. Bawendi, J. Phys. Chem. B 2002, 106, 7619 7622. [14] J. K. Jaiswal, H. Mattoussi, J. M. Mauro, S. M. Simon, Nat. Biotechnol. 2003, 21, 47 51. [15] C. R. Kagan, C. B. Murray, M. Nirmal, M. G. Bawendi, Phys. Rev. Lett. 1996, 76, 1517 1520.

ChemPhysChem 2006, 7, 47 57

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemphyschem.org

57

Vous aimerez peut-être aussi