Exploring attachment chemistry with FRET in hybrid quantum dot dye-labeled DNA dendrimer composites

Anirban Samanta ab, Susan Buckhout-White a, Eunkeu Oh cd, Kimihiro Susumu cd and Igor L. Medintz *a
aCenter for Bio/Molecular Science and Engineering, Code 6900, U.S. Naval Research Laboratory, Washington, DC 20375, USA. E-mail: Igor.medintz@nrl.navy.mil
bCollege of Science, George Mason University, Fairfax, VA 22030, USA
cOptical Sciences Division, Code 5600, U.S. Naval Research Laboratory, Washington, DC 20375, USA
dKeyW Corporation, Hanover, MD 21076, USA

Received 20th November 2017 , Accepted 21st December 2017

First published on 5th January 2018

Luminescent semiconductor quantum dots (QDs) and a range of biomolecules are now being routinely co-integrated into functional optical devices in pursuit of creating novel ‘value added’ photonic and energy harvesting/transfer materials. Amongst the biological molecules, structural DNA architectures are particularly useful due to their unrivaled ability to assume almost any desired shape along with allowing fluorophores to be precisely arranged on them with controlled stoichiometry and sub-nanometer positional accuracy. The unique properties available to joint QD–DNA composites suggest them for a host of new applications in light harvesting, biosensing, and molecular computation amongst others. To fully realize the synergistic benefits from such organic–inorganic composites, especially when they constitute complex, multidimensional Förster resonance energy transfer (FRET) networks, a detailed understanding of the mechanisms that govern the individual components is imperative. Here, we demonstrate hybrid FRET systems comprising an initial QD scaffold/donor displaying DNA dendrimers decorated with dyes and which are capable of efficiently capturing UV light and transporting it to spectrally and spatially distant fluorophores via multistep FRET. We evaluate two primary strategies to conjugate the DNA-dendrimers to the QDs, namely covalent attachment of DNA to the termini of the QDs surface ligands and polyhistidine-based metal affinity coordination of modified DNA to the QD's ZnS shell surface. Analysis of the resulting FRET data shows that the dendritic arrangement of the dyes and the ability to place multiple dendrimer copies around the QD's nontrivial surface provides for significant energy transfer efficiencies of 20–25% through these multi-FRET step systems. In analyzing the properties of the conjugates, we further find that each assembly chemistry brings with it a series of benefits and liabilities that serve as mutual trade-offs and potential rules of thumb for designing future nanodevices based on these materials.

Design, System, Application

Integrating nanomaterials with biological molecules will allow the creation of de novo systems which functionally exploit the unique properties available to both components. Such molecular engineering, however, is almost exclusively dependent on the assembly chemistries utilized to integrate both components and this should ideally occur with intimate control over all dependent properties. Here, we pursue light harvesting and energy directing systems comprised of a central luminescent semiconductor quantum dot (QD) displaying multiple pendant DNA dendrimers which, in turn, are decorated with organic dyes. We focus on evaluating two assembly chemistries to create these hybrid organic–inorganic systems; covalent attachment of DNA to the QD surface ligand and metal affinity coordination of modified DNA to the QD shell. Analyzing the photophysical properties and energy transfer efficiency within these systems reveals that each chemistry is associated with its own benefits and liabilities and these can be used as potential rules of thumb for designing future nanodevices based on similar architectures. Insight provided from these lessons will help not only create an improved generation of these devices but is also applicable to other engineered nanoscale devices that seek to integrate disparate biological and nanomaterial components.

1. Introduction

Significant research efforts are currently focused towards developing new ways to manipulate optical energy at the nanoscale level. As part of these efforts, a range of source materials including inorganic, organic, and even those that are biological in nature are being explored to generate a library of de novo nanoscopic optically-active devices with potential applications in diverse areas ranging from optical data storage, logic and computation, to biosensing and even artificial photosynthesis.1,2 Molecular photonic wires (MPWs) are of particular interest in this arena since they function as subwavelength-sized waveguides by transporting excited electronic energy in the form of optical signal from an input to an output unit and can be useful in molecular circuitry that suffers from a “physical contact” issue. The general outline of building MPWs is to create a regular arrangement of dyes, either chemically attached in a linear fashion or closely assembled on polymeric or biological scaffolds so that near field or other optical/electronic interactions are possible.3–6 In these weakly coupled systems, judiciously selected fluorescent dyes with linearly arranged absorption and emission profiles allow for a downhill cascade of excited state energy from the input to output unit primarily via a Förster resonance energy transfer (FRET) mechanism.7,8

Amongst nanoscale scaffolding and assembly materials, DNA has turned out to be very useful for building a variety of nanoscale optical devices including especially MPWs.9–14 They can be designed in silico and virtually any conceivable 1-, 2-, or 3-dimensional nanoarchitecture with impressive curvature can be constructed using DNA as the sole building material. Besides the highly predictable Watson–Crick base complimentarily that drives the self-assembly of these DNA nanostructures, easy synthesis, amenability to a wide variety of fluorophore modifications, enzymatic labeling and most importantly sub-nanometer positional accuracy has made them an extremely potent material for serving many purposes including optical energy manipulation in the nanometer regime.

Several DNA-scaffolded MPW designs have been recently described both on the ensemble and single molecule level, helping to shed light on their inherently complex photophysical properties while also seeking to improve their exciton delivery efficacy.3–6,15 In seminal work, Hannestad et al. devised a MPW with covalently attached dyes as injector and detector at the two termini, respectively, and incorporated non-specific intermediary or relay intercalating dyes that participated in homoFRET to enable energy transfer (ET) over a distance of 20 nm.16 Diaz studied a similar system utilizing covalently-tethered bridging fluorophores, which not only extended the overall exciton transport distance but also helped improve its end-to-end exciton transfer efficiency.17 Both Spillmann et al. and Sanchez-Mosteiro et al. investigated the cause of lower than expected exciton delivery efficiency values in multichromophoric DNA photonic wires.18,19 They concluded that multiple factors including poor dye performance due to local environmental effects, incomplete hybridization, and unfavorable dipole orientation contributed to the low performance. Using a systematic structure–function approach, Buckhout-White et al. showed that use of multiple, overlapping FRET pathways such as those found in dendrimeric DNA structures could increase the ET efficiency considerably within MPWs.20 Moreover, the dendrimer's multiple branches could provide inherent redundancy to help address any assembly or dye incorporation deficiencies. The benefits that dendrimeric MPWs offer have been recently extended to include bioluminescent enzymes as an intrinsic light source.21

While a rich library of fluorophores are now available for bioconjugation and incorporation into biological scaffolds, luminescent semiconductor nanocrystals or quantum dots (QDs) offer unrivaled advantages as an initial FRET sensitizer that are cumulatively unattainable to most dyes.22–32 Their unique quantum confined optical properties include broad absorption coupled to narrow-nearly symmetrical photoluminescence (PL) profiles allowing for excitation with minimal contributions to any downstream chromophores present in the system, unusually high resistance to photobleaching, chemical stability, and high quantum yield (QYs). Besides these unique photophysical properties, their large surface area and surface-to-volume ratios allow for attachment and display of multiple surrounding acceptor chromophores in a centrosymmetrical manner.22,23 This can proportionally increase the acceptor absorption cross-section and allow controlled tuning of subsequent FRET efficiency within the conjugate. We previously demonstrated several generations of QD donor-assembled multi-acceptor MPWs that manifested a multistep FRET cascade delivering excitonic energy to a spatially and spectrally distant terminal acceptor dye.33,34 Here, the QD provides the two key roles of initial light harvesting/sensitizer along with that of assembly platform within these structures. In these systems, the end-to-end exciton transfer efficiency was improved from an initial <0.1% in the first iteration to nearly 10% in the second by optimizing dye pairings, engineering dye spacing as well as wire display valency around the QD.33,34 However, the MPW portion of the structure was essentially linear and 1-dimensional which limited exciton transport. Imbuing these types of composite photonic structures with multiple, redundant exciton pathways could certainly help to increase ET efficiency while enhancing their complexity.

The process of functionally integrating DNA dendrimers with QDs begins with a key fundamental step – namely the bioconjugation chemistry that physically links the DNA components to the central QD scaffold. Ideally, this should be performed in a manner that allows for careful control over all pertinent structural and functional variables.35 Here, we evaluate two primary QD-bioconjugation chemistries for these specific purposes, see Fig. 1. The first chemistry, referred to as strategy I, covalently attaches amine-modified DNA to the terminal carboxyl groups displayed on the QD's surface functionalizing ligand. Strategy II coordinates peptide modified DNA directly to the QDs shell. Each of the resulting conjugates were then subjected to an initial photophysical analysis to evaluate the resulting performance. Subsequent analysis indicates that each approach brings with it a series of benefits and liabilities that may need to be carefully considered when creating such composite QD-MPW structures.

image file: c7me00121e-f1.tif
Fig. 1 Schematic of the QD-dendrimer structures and linkage chemistry. (A) Schematic depicting a concatenated 2-to-1 DNA dendrimer decorated with pendant fluorophores affixed to a QD with an emission maxima at 540 nm. The circles illustrate the dyes at their estimated locations while each of the DNA strands are presented with distinct colors. For clarity only one dendrimer is shown appended to the QD, but ideally there should be six of them centrosymmetrically arrayed around it. Excitation of the QDs at 400 nm triggers a 4-step FRET cascade from the central QD to the dye acceptors that were judiciously chosen to facilitate transport of energy across the visible to the near infrared region of spectrum. (B) Schematic illustrating the chemical structure of the linkers connecting the DNA dendrimers to the QDs. (i) Strategy I. Using EDC coupling chemistry, an amide bond is created between the terminal primary amine on modified DNA and a carboxylic acid group on the DHLA-NTA bearing capping ligands on the QD. (ii) Strategy II, Hisn-metal affinity coordination chemistry was used to conjugate the DNA to the CdSe/CdZnS/ZnS core/shell/shell QDs. A His5AlaCys segment was covalently attached to amine modified DNA via a heterobifunctional SMCC cross linker. One histidine is shown unbound solely for 2D drawing purpose. (iii) Sequences of the dye-labeled DNA strands attached to the QDs and their complimentary pairs. The purple colored strand has the same sequence as the yellow one, just without any dye modification. Note, schematics are not to scale.

2. Materials and methods

2.1. Quantum dots

The 540 nm emitting CdSe/CdZnS/ZnS core/shell/shell QDs utilized were synthesized as described.36,37 The QDs were made hydrophilic by exchanging the native hydrophobic capping ligands with dihydrolipoic acid modified nitrilotriacetic acid (DHLA-NTA), see ESI Fig. S1 for the structure.38 Following cap exchange with this ligand, the QD dispersions were found to be highly stable and bright for prolonged periods of time when stored at 4 °C. Structural characterization of the as-prepared QDs was carried out using a JEOL JEM-2100 FE-TEM, field-emission gun transmission electron microscope, providing high spatial resolution atomic imaging and microstructure analysis of material samples. Samples for TEM were prepared by spreading a 10 μl drop of the filtered QD dispersion (filtered using 0.25 μm Millipore syringe filters) onto ultrathin carbon/holey support film on a 300 mesh Au grid (Ted Pella, Inc.) and letting it dry. The concentration of QDs used for TEM was typically ∼1 μM. Individual particle sizes were measured using a Gatan Digital Micrograph (Pleasanton, CA); average sizes along with standard deviations were extracted from analysis of ∼100 QDs.

2.2. DNA and dendrimer assembly

DNA sequences were purchased from IDT, Inc. with the exception of the Cy3.5-labeled strand which was purchased from Operon Biotechnologies, Inc. DNA sequences and dye insertion locations can be found in the ESI. Depending on the design requirement, the dyes were incorporated into the DNA strands either terminally or internally. For example, the final Cy5.5 acceptor dye situated at the dendrimer's outer periphery was terminally attached to its oligonucleotide via a C6 alkyl chain. The long alkyl chain introduced some flexibility in dye position and orientation. The internally modified dyes with the exception of Cy3.5 were incorporated directly as phosphoramidites and are expected to be far more rigidly placed with two attachment points and are assumed to be oriented parallel to the DNA backbone.39 For the Cy3.5 dye, succinimidyl ester conjugation chemistry to a modified T* (T* = amino C6 dT) base with an amine was used which introduced a C6 alkyl chain between the dye and the DNA backbone again allowing for some positional and orientational movement.

For metal-affinity coordination to QDs, a peptide-DNA hybrid was utilized similar to that described previously.40 This consisted of the sequence ACTCCCTGTATCCCGTGACTTAACTCGTGAACTCGTGT-linker-Cys-Ala-(His)5, where the heterobifunctional linker succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC) connects a 5′ C6-amine group on the DNA to the peptide's cysteine thiol. Dendrimers were typically assembled by mixing the oligos, unmodified and dye-modified as per experimental requirement, at appropriate ratios in 2.5× PBS buffer (phosphate buffered saline: 2.5 mM phosphate buffer, 6.8 mM KCl and 343 mM NaCl ∼pH 7.4). The mixture was heated at 90 °C for a brief period of time and subsequently cooled down to room temperature over the course of 12 hours using a PCR thermal cycler. Considering their potential downstream use, they were used for photophysical measurements without further purification.

2.3. QD–DNA conjugation chemistry

Two ligation strategies, covalent conjugation and polyhistidine-(Hisn) metal affinity coordination, were utilized in the present study to affix the DNA dendrimers to the QDs.41 In strategy I, standard 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC)/N-hydroxysulfosuccinimide (sulfo-NHS) coupling chemistry was exploited to covalently link primary amine labeled DNA to the carboxylic acid groups displayed on the QD surface.41–43 In typical reactions, 10 pmol of QDs were reacted with 10[thin space (1/6-em)]000 fold excess of EDC in the presence of 500 fold excess sulfo-NHS. This huge excess of the activating agent was used due to its rapid hydrolysis in water.44 The reaction was performed in 2.5× PBS buffer at pH 7.4. Due to the semi-stable nature of the intermediate, a 20 fold excess of the amine-DNA was quickly added to the mixture and left for an hour under mild agitation. After passing through a desalting PD-10 column (GE Healthcare) to get rid of the EDC and NHS, the reaction mixture was concentrated under vacuum and purified through a P = 60 polyacrylamide resin column (Biorad) to get rid of any excess unreacted DNA that was present. The DNA-dendrimer, with a protruding sticky end designed for hybridizing with the DNA displayed on the QD, was mixed with the DNA conjugated QD at a ratio of six dendrimer to one QD (0.25 μM) since each QD could accommodate more than this (typically >10/QD as determined from analyzing the DNA remaining in the reaction mixture and flow through). To facilitate hybridization, the resulting mixture was heated at 40 °C for a brief period of time and subsequently allowed to cool down to room temperature over the course of an hour. Strategy II achieves conjugation by allowing Hisn-modified DNA to coordinate directly to the QD's ZnS surface.41,45 Preformed DNA dendrimer bearing the His6 tag pendant on one of the constituent oligos was mixed with the QD maintaining the same ratio and concentration as above, 6 dendrimer to 1 QD (0.25 uM) in 1× PBS buffer, and left in the dark at room temperature for an hour to assemble.

2.4. Photophysical analysis

Steady state fluorescence spectra were collected at room temperature on a NanoLog spectrometer (HORIBA Jobin Yvon) equipped with a thermoelectric cooled PMT (R928 in the range 200 to 850 nm). Samples were excited at 400 nm excitation as this is close to the QD's second absorption band while also corresponding to dye minima (Fig. 2A); this helps ensure that only the QD donor is primarily excited. To quantify and compare performance of the FRET steps in each structure as well as overall exciton transfer efficiencies, photon contributions from the QD and individual sensitized dyes are needed. After accounting for instrumental effects, the composite spectra were deconstructed into individual fluorophore components, as described.15,18,20,34 From the steady state data, the average FRET efficiency E was estimated using the following expression:
image file: c7me00121e-t1.tif(1)
here, FD and FDA are the integrated fluorescence intensity of donor in the absence and presence of the acceptor, respectively. Donor–acceptor separation (r) in the case of a single donor and multiple acceptors, can be estimated using:
image file: c7me00121e-t2.tif(2)
where n designates the number of acceptors and R0 is the Förster distance (donor acceptor separation where FRET E = 50%). R0 is typically calculated with:
R06 = 0.0212ΦDn−4(in nm6)(3)

image file: c7me00121e-f2.tif
Fig. 2 Photophysical properties of relevant fluorophores. (A) Normalized absorption/emission spectra of the QD and four dyes used. (B) Spectral overlap integrands versus wavelength for sequential donor–acceptor combinations. (C) Spectral overlap integrands versus wavelength for the initial QD donor and each of the sequential acceptor dyes to evaluate the possibility of long-range energy transfer in the structures.

Here J is the donor emission–acceptor absorption overlap integral, κ2 is the dipole orientation factor (typically 2/3 for the random orientations present in these self-assembled structures), ΦD is the donor QY, and n is the refractive index of the media.7,8

Overall energy transfer efficiency and FRET network performance were also evaluated in terms of anywhere-to-end exciton transfer efficiency (Eae) using the empirical formula as outlined previously.16,18,20,21,34 This estimates the probability of photon transfer from the initial donor or any other dye input point to the terminal acceptor and is expressed as:

image file: c7me00121e-t3.tif(4)
where ϕA and ϕD are the QY of the final acceptor and initial donor, respectively, FD denotes the integrated fluorescence intensity of the donor in the absence of the acceptor, FAD denotes the integrated fluorescence intensity of the final acceptor in the presence of the donor (as well as all intermediate dyes), and FA is the fluorescence intensity of the final acceptor with all intermediate dyes but not the initial donor. This metric was determined directly from the spectra of the full constructs assembled as indicated above in comparison to the relevant controls. In contrast to our previous use of an end-to-end transfer metric,33,34Eae helps account for all manner of excitation and provides a good metric of overall energy transfer to the terminal acceptor in such systems. We thus rely on the latter descriptor here and it is functionally compared to the end-to-end values reported in previous constructs. All other FRET parameters were determined using standard FRET formalism.8

3. Results and discussion

3.1. DNA dendrimer and spectral overlap

The primary goal of this study was to create a hybrid multi-step FRET QD–DNA dendrimer MPW assembly and to specifically investigate how attachment chemistry might influence the efficiency of the resulting FRET cascade. The DNA dendrimer design was inspired from Luo's original report of creating dendritic structures with Y-shaped DNA tiles and slightly modified as described previously.20,46 Instead of concentrating energy from the periphery to the core, here the photonic energy is intended to funnel out from the QD center to the periphery via multistep FRET, as depicted in Fig. 1A. Both directionalities have been previously investigated and utilized in different functional variants of these dendrimeric structures.20,21

The primary modification to the previously reported structure was a nick introduced in the central strand and here the loose ends were extended by a few nucleotides to create a QD attachment site, see Fig. 1A.20 The branching motif is designated as F1 → [F2 → [F3 → [F4 → F5n]n]n]n (→ = FRET step); where F1 represents the first fluorophore donor, the QD in this case, and n stands for the branching ratio and reflects the number of sequential acceptors fed by each dye when that dye is in the role of a relay/donor. A higher acceptor redundancy, i.e. a higher branching ratio, could help increase overall ET through the system, however, the ensuing structural complexity and associated issues of formation and yield resulted in us selecting a branching ratio of 2. Although this design is structurally and functionally analogous to that utilized previously,20 actual sequences were updated and re-optimized. Structures were visualized in the coarse-grain modeling program Nanoengineer which also identifies problematic junctions having excessive distortions and suggests alternate sequences. Dye placement sites were chosen within the dendrimers to keep individual donor–acceptor inter-dye spacing's as close as possible to ≤0.5× the Förster distance (R0); here the predicted FRET E at each ET step should be ≥90%. Predicted distances between the donor and acceptor were calculated taking into account several factors including the number of nucleobases separating them, their relative orientation on the double helix, and the alkyl chain linking the dye to the DNA bases/backbones.20,34 These donor–acceptor distances are also much shorter than the double stranded (ds) DNA persistence length and hence some of the spacing's and portions of the dendrimeric structure could be assumed to be rigid. DNA sequences along with the corresponding melting temperature are reported in the ESI.

The intended functionality of the FRET cascade requires that the QD be excited as the initial donor and then transfer energy to the first, proximal dye acceptor on the DNA dendrimer. Now in the role of an intermediary or relay, this first fluorophore then transfers the exciton energy to the next acceptor dye which, in turn, then becomes a similar relay and this process is propagated stepwise through the rest of the structure to the terminal acceptor. This FRET cascade process is predicated on the constituent fluorophores displaying the requisite donor–acceptor properties to facilitate each intended step. Fig. 2A plots the absorption and emission of the QD and each sequential relay dye while Fig. 2B plots the integrand of the spectral overlap function versus wavelength for each single donor–acceptor interaction. Fig. 2C shows the same information for the initial QD donor with each of the individual dyes, this is meant to highlight the possibility for longer range ET interactions. Such longer range FRET interactions between fluorophores that are spectrally and spatially separated such as QD → Cy5, for example, should be insignificant unless the dyes are brought into very close proximity or there are multiple Cy5 dyes arrayed around the QD. Table 1 lists relevant photophysical properties of this system including each fluorophore's QY, absorption and PL maxima, extinction coefficients, estimated R0 values and spectral overlap functions, J. Since there is a non-trivial probability for homoFRET interactions between the same dyes in this dendritic arrangement, each dye's spectral overlap with itself is also listed. Assuming a dynamic isotropic limit shows that the Förster distance varies quite widely from ∼34 Å (QD → Cy5.5) to ∼67 Å (Cy5 → Cy5.5) whereas the J values vary by orders of magnitude from 1.31 × 10−12 (Cy3.5 → Cy5.5) to 8.54 × 10−14 (QD → Cy5.5). Overall, this data confirms that the necessary requirements for a sequential FRET cascade are in place due to the existence of contiguous spectral overlap from the primary donor QD to the final acceptor Cy5.5. This slanted energy landscape should allow a downhill cascade of ET steps from the initial QD to the terminal Cy5.5 dye across a wide portion of the spectral range and significant spatial distance.

Table 1 Relevant fluorophore photophysical properties
Fluorophore QY (%) Ext. coeff. (M−1 cm−1) λ max abs. (nm) λ max em. (nm) J in cm3 M−1 (R0 in Å)c
540 QD Cy3 Cy3.5 Cy5 Cy5.5
a Chemical strategy. b QD extinction coefficient is reported at 350 nm which corresponds with the second absorption band. c R 0 calculations utilize n = 1.33, and the assumption of κ2 = 2/3.7
540 QD (I)a 0.12 241[thin space (1/6-em)]110b 540 6.06 × 10−13 (47) 9.27 × 10−13 (51) 7.41 × 10−13 (49) 1.59 × 10−13 (38) 8.54 × 10−14 (34)
540 QD (II) 0.32 241[thin space (1/6-em)]110b 540 6.06 × 10−13 (56) 9.27 × 10−13 (60) 7.41 × 10−13 (58) 1.59 × 10−13 (45) 8.54 × 10−14 (40)
Cy3 0.15 150[thin space (1/6-em)]000 548 565 3.21 × 10−13 (44) 9.02 × 10−13 (52) 8.86 × 10−13 (52) 5.05 × 10−13 (48)
Cy3.5 0.15 150[thin space (1/6-em)]000 591 605 3.62 × 10−13 (45) 1.63 × 10−12 (58) 1.31 × 10−12 (56)
Cy5 0.28 250[thin space (1/6-em)]000 648 668 1.56 × 10−12 (64) 2.19 × 10−12 (67)
Cy5.5 0.23 190[thin space (1/6-em)]000 688 705 2.23 × 10−12 (65)

3.2. Conjugation chemistries

Strategy I covalently attaches amine-modified DNA to the carboxyl groups displayed on the QD's surface functionalizing ligand using a zero-length carbodiimide linker. QDs surface functionalized with DHLA-NTA ligands were utilized as the common scaffold in this study.36,38 Although the NTA ligand can coordinate divalent metal cations and the subsequently “charged” NTA group can participate in metal affinity coordination directly, it was not utilized here in this role to avoid any metal ion quenching of QD or fluorophore PL.41 Rather, the triacetic acid groups served to provide as many carboxyl groups as possible for covalent conjugation to the aminolated-DNA. EDC-based chemistry was the original approach utilized in one of the first reports of QD application as a bioprobe.47 Along with streptavidin–biotin conjugation, EDC chemistry still remains one of the most popular QD biomodification chemistries although it can be complicated depending upon which ligand type is utilized.48 For example, QDs surface-stabilized with mercaptoalkyl carboxylic acid ligands (e.g., mercaptoundecanoic acid) tend to irreversibly lose colloidal stability during EDC modification reactions due to the rapid formation of a poorly soluble O-acyl urea intermediary.44 Moreover, the presence of multiple amines and carboxyls on proteins can induce rapid QD-protein crosslinking necessitating careful control over stoichiometry in the conjugation reaction. Limiting the reaction pathways and ability to crosslink by including only a single amine-modified DNA in the mix along with use of more stable zwitterionic ligands displaying tertiary amines and multiple carboxyl groups such as the current DHLA-NTA has been shown to provide for viable QD conjugates while negating crosslinking when implementing EDC chemistry.36 A similar approach is utilized here by only providing a single amine on the DNA as target for the activated QD surface. EDC conjugation of aminolated DNA to the QDs was carried out as described in the Materials and methods using a 20× molar excess of DNA per QD. Following purification, this typically yielded an average of 10 or more DNA oligo's attached per QD indicating a ≥50% coupling efficiency (data not shown). For final conjugate assembly, the DNA dendrimer equipped with a sticky end which was complimentary to the covalently affixed DNA on the QD were rapidly hybridized as described in the Methods. A ratio of 6 DNA dendrimers per QD was used during hybridization to minimize issues associated with Poisson distribution effects that are typically found when using assembly ratios of <3–4 moieties/QD during a conjugation reaction.41Fig. 1Bi schematically depicts the chemical structure of this final conjugate.

Strategy II relies on metal affinity coordination between adjacent Hisn residues and the QD surface. Here, the imidazole side chain groups on the histidine bind strongly to the QD's Zn rich surface. QD bioconjugation based on this high affinity interaction (Kd ∼ 1 × 10−9 M) has been repeatedly proven by a growing number of groups with a plethora of His6-labeled proteins and peptides.41 More pertinently, it has been repeatedly utilized to attach Hisn-modified DNA fragments to QDs.38,40,49–53 The primary benefit of this QD conjugation approach is its simplicity along with the relative control it provides over average display valency and biomolecular display orientation on the QD.41 Such QD conjugates almost instantaneously self-assemble following a mixing of stoichiometric quantities of each participant and are ready for use in most cases without requiring any subsequent purification steps.54–56 The long-term stability of QD-conjugates assembled in this manner has been repeatedly verified in a number of complex biological matrices and animal systems.57,58 To facilitate this approach, the chimeric peptido-DNA ACTCCCTGTATCCCGTGACTTAACTCGTGAACTCGTGT-linker-Cys-Ala-His5 was synthesized. In this construct, the terminal amine on the DNA is linked to the thiol on the peptide portion through an SMCC heterobifunctional linker. Fig. 1Bii schematically depicts this linkage between the histidine-modified DNA and the QD. To allow for a one-to-one comparison between the two QD-dendrimer systems, the DHLA-NTA functionalized QDs were again utilized here but without any loading of metal cations into the NTA so that the Hisn portion binds directly to the QD surface. It is also important to note that Hisn-Zn coordination to the QD surface does not displace the surface ligands present but rather binds to available sites presumably around them as dictated by steric considerations.59 For final conjugate formation, the full DNA dendrimer structure was pre-assembled by hybridization and then subsequently coordinated to the QDs at a 6[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio for parity with the previous approach.

One of the critical elements arising directly from the conjugation chemistry that could potentially affect FRET E is that of the QD-dendrimer separation distance. This will have the most pronounced effect on the very first FRET step (QD → Cy3) and can, in turn, affect the efficiency of subsequent ET steps in the cascade. The C6 alkane linker that connects the terminal amine to the rest of the DNA sequence is common in both chemistries. In strategy I, the DNA is attached to the DHLA-NTA ligand at its termini by this linker and this should contribute around ∼2.2 nm of extension from the QD surface in a fully extended conformation.36 In contrast, the linker in the His6-peptide-DNA construct consists of the alanine and cysteine residues, the SMCC moiety and the C6 alkane. However, since the Hisn directly coordinates to the Zn on the QD surface, the distance from the QD surface to the beginning of the DNA segment should not be appreciably different between the two chemical linkage strategies.

3.3. Verification of QD–DNA dendrimer MPW conjugate formation

Prior to analyzing FRET in the hybrid QD–DNA MPWs, it was important to first confirm that the structures were indeed forming as desired. Fig. 3A shows some representative TEM micrographs of the as-prepared QDs with a diameter of 4.6 ± 0.4 nm. Fig. 3B shows a representative image collected from separating the fully hybridized DNA dendrimer product as assembled with unlabeled oligos by polyacrylamide gel electrophoresis (PAGE) in parallel with a 50 base pair (bp) DNA ladder. The resulting gel was stained and visualized with gel red staining. The intense band corresponds to the desired fully-formed dendrimer while the faint trail signifies some concatenated assemblies, partial structures and unbound oligos which are typically present within these types of dynamic hybridization-driven structures. The dendrimer migrates at a molecular mass that corresponds to a dsDNA of ∼300–325 bp. The estimated dendrimer size based just on the number of bps of DNA present is 16 nm × 11 nm with a total of ∼106 bp. The difference in migration is ascribed to the complexity of the dendrimer's 3-D DNA structure. Predicting the migration size of such complex 3-D structures during electrophoresis is a capability that is still not fully realized.60 More pertinently, comparison of the integrated PL of the isolated main band versus that of the entire gel lane allows us to estimate a formation yield of ∼80%. This is actually much higher than that attained previously with similar types of dendrimeric structures due to sequence optimization.20,21
image file: c7me00121e-f3.tif
Fig. 3 QD structure and gel analysis of DNA constructs. (A) Representative TEM micrographs of the QDs with a diameter of 4.6 ± 0.4 nm. Inset shows a high-resolution image of individual QDs highlighting the crystalline lattice structure. TEM data was collected and analyzed as described in the Material and methods and ref. 73. (Bi) Native polyacrylamide gel electrophoresis of the DNA dendrimer with gel red staining. From comparison to the 50 bp DNA ladder, it can be concluded that the intense distinct band is the desired product while the smeared band appears to be some partially formed structure and/or simply unbound strands. Agarose gel electrophoretic separation of QD–DNA dendrimer conjugates assembled with (ii) strategy I – EDC chemistry or (iii) strategy II – Hisn-metal affinity coordination. Increasing ratios of dendrimer to QD as noted. For the gel in panel ii, the 0 sample reflects the QD following covalent modification with DNA that is complementary to the dendrimer. Higher number of dendrimers makes the composite structure larger and more charged resulting in a change in migration that is visibly different for each strategy. Both gels were run using 1.5% agarose in 1× Tris-borate-EDTA buffer (TBE, 89 mM Tris-borate and 2 mM EDTA, pH 8.3) at 10 V cm−1 for 20 min.

Composite QD-dendrimer structures assembled using both chemical approaches were also analyzed by separation in a 1.5% weight/volume agarose gel supplemented with 1× Tris-borate-EDTA buffer. A representative gel of QDs with increasing ratios of dendrimer prepared using the covalent EDC-based strategy I is shown in Fig. 3Bii. The QD “0” sample contains QD with DNA complementary to the dendrimer covalently attached to the surface but no dendrimer hybridized to it. Here, the migration rate slows in a manner that tracks proportionally with the increasing ratio of dendrimer attached to the QDs surface. Fig. 3Biii presents a representative image of QDs assembled with an increasing ratio of dendrimer via the second Hisn-based metal coordination approach. The QD “0” sample in this case corresponds to a QD without anything on its surface or chemical modification to its ligand. Here, too, the QD's migration is also slowed when the dendrimer is first attached to the QD surface at low valency. However, an increasing ratio of dendrimeric DNA with its high intrinsic negative charge now increases the QD-composite's migration rate in contrast to what is seen just above it. Such differences in the manner in which the QD–DNA composite migrates in a gel are also hard to predict a priori and reflect the overall mass/charge ratio which, in turn, reflects the complex interplay between QD surface ligand charge, DNA charge, QD colloid and DNA mass, choice of conjugation chemistry implemented, choice of sieving media, buffer choice, and, indeed, such migration differences have been noted before.49,50,61 The QD's inherent surface ligand charge will be altered by the covalent approach as well as being the attachment point for a charged DNA oligo while it will remain unperturbed with the metal affinity attachment approach. The slightly brighter QD bands in the composite structures are believed to reflect better surface passivation effects following bioconjugation and have also been noted quite frequently before as well as sometimes serving as a confirmation of bioconjugation in its own right.41 Nevertheless, both gels serve to confirm that the dendrimer is indeed attaching to the QD in a ratiometric manner by both conjugation approaches as expected.

3.4. FRET in QD-dendrimer MPWs assembled by strategy I – covalent chemistry

FRET progression in various QD-dendrimer MPW constructs prepared via covalent conjugation were assessed first. All ds dendrimers were assembled with the full complement of DNA while substituting unlabeled oligos for acceptor-dye labeled variants as experimentally required. Fig. 4A shows the spectral evolution as each sequential acceptor dye is incorporated into the QD-MPW construct and compared to the previous. Addition of the first 12 surrounding Cy3 acceptor dyes (6 dendrimers × 2 Cy3 each) around the QD donor decreases its PL ∼50%. QD donor PL continues to decrease as successive acceptor dyes are added in the dendrimer with concomitant terminal dye acceptor sensitization clearly apparent from just the raw spectra. The continuing decreases in QD PL with addition of each acceptor dye suggests some direct interactions between the QD and those additional, more distantly placed acceptors. Interestingly, the Cy3 and Cy3.5 emission in constructs containing downstream acceptors, e.g. QD → Cy3 → Cy3.5 → Cy5, is almost completely gone suggesting that they may be functioning as relays with unity efficiency. Similar findings for these dyes have been noted before in QDs displaying linear DNA wires.34 Analyzing the PL spectra from the simplest QD–Cy3 construct following spectral deconstruction suggested that although the FRET E was significant at ∼53% this value was still slightly lower than our previous findings of >60% with similar dye display valency around the QD.34 The high melting temperature (Tm ∼ 76 °C) of the dendrimer sticky end that binds to the QD-affixed DNA suggests an incomplete hybridization could be ruled out. Reexamining the QD QY revealed that it had deteriorated substantially from 0.32 to 0.12 or by ∼63%. Control experiments showed chemical modification with EDC and the multistep purification thereafter to be the source of this decrease. It is believed that the large excess of highly reactive EDC deleteriously impacted the QD's surface passivation and hence it's QY in this case. Table I presents the adjusted R0 values considering the new QD QY value. The Förster-weighted separation distance (rDA) between the QD center and Cy3 dye incorporating this new QY value was estimated at ∼76 Å, similar to that seen in a prior study with a similar architecture and in agreement with the QD diameter.34 Cy3 acceptor sensitization was estimated at ∼40% suggesting the existence of either some other non-radiative relaxation pathways or, alternatively, quenching effects from a nearby nucleobase or even complex homoFRET interactions between adjacent Cy3 dyes.17
image file: c7me00121e-f4.tif
Fig. 4 Representative plots showing FRET evolution and progression in the QD–DNA dendrimer constructs prepared via strategy I – EDC-based covalent conjugation. (A) PL spectra collected as sequential acceptor dyes were incorporated in the dendrimer. As denoted, the black trace corresponds to the emission spectrum of QDs while others show clear signature of the FRET sensitized dye components. The QD only sample is assembled with unlabeled dendrimeric DNA. (B) Spectral evolution of control samples with one missing dye plotted as superimposed over the construct with maximum dye display (red trace). A significant degree of FRET takes place despite interruption in the sequential FRET. (C) Spectra of another set of controls with two selected dyes missing. (D) Dotted traces displaying the PL spectra collected from QD-dendrimer constructs with only the one indicated acceptor dye present while the continuous traces show the intensity profiles of the directly excited dyes at the same excitation wavelength.

Along with the sequentially constructed configuration presented in Fig. 4A, other relevant control constructs with one or more missing dyes were also interrogated to understand the internal exciton transfer dynamics (Fig. 4B–D). Table 2 presents the estimated donor quenching and Eae as a percentage value along with the rDA values for QD separation to the remaining Cy3.5, Cy5, and Cy5.5 dyes as determined from control constructs where only each acceptor dye was present in the dendrimers around the QD (Fig. 4D). The calculations used to derive these estimates do account for the presence of multiple copies of each acceptor around the central QD donor. The latter show that the Cy3.5 and Cy5 dyes are further away from the QD as expected by ca. 3–4 Å respectively. Interestingly, the terminal Cy5.5 dye demonstrates an average rDA of 75 Å suggesting that it is closer to the QD than the first Cy3 acceptor/relay dye. It is currently not clear if this anomalous result arises from an actual bending of the dendrimer structure back towards the QD or is an unexpected effect arising from placement of ∼48 Cy5.5 dyes around the central QD with each displaying a high 190[thin space (1/6-em)]000 M−1 cm−1 extinction coefficient; the latter gives rise to a total acceptor extinction coefficient of >9.1 × 106 M−1 cm−1. A high FRET E over a significantly extended rDA was similarly noted for QDs decorated with the beta-phycoerythrin light harvesting complex which contains ∼34 open-chain tetrapyrrole bilin chromophores displaying a cumulative acceptor extinction coefficient of >2.4 × 106 M−1 cm−1.62 We currently hypothesize that the dendrimer's 3-dimensional shape and rotational flexibility in its arms serves to place some Cy5.5 acceptor dyes closer to the QD donor surface. These closer Cy5.5s are far more favored to dominate the FRET interactions with the QD by virtue of their proximity and thus give rise to the shortened separation.8 The fact that the QD is quenched by ∼25% when only surrounded by dendrimer labeled with Cy5.5 supports this along with a direct QD → Cy5.5 Eae of 8%. Another possibility which cannot be discounted is that the large number of Cy5.5 acceptors displayed around each QD donor may cumulatively act to “skew” the single donor–single acceptor 6th power FRET distance dependency to a somewhat lower value; this plasticity has been postulated before but remains hard to prove unequivocally in a defined experimental format.63 The fact that this is not seen with the 24 surrounding Cy5 dyes with their even closer distance to the QD donor and higher extinction coefficient argues against it, however.

Table 2 Estimated donor quenching and Eae in selected QD-dendrimer configurations
Construct: QD Cy3 Cy3.5 Cy5 Cy5.5
Notes. All values have standard deviations less than 5%, not shown for simplicity. X indicates dye replaced with DNA. Eae is anywhere-to-end energy transfer efficiency.15,20 QD to dye rDA is assumed to be from the QD center to the dye center.8,22
Strategy I. Covalent EDC-based chemistry Donor loss % [Eae%]
QD–Cy3 53 [53]
QD–Cy3–Cy3.5 67 100 [53]
QD–Cy3–Cy3.5–Cy5 69 100 85 [30]
QD–Cy3–Cy3.5–Cy5–Cy5.5 79 100 92 23 [20]
With missing dyes
QD–X–Cy3.5 41 [44]
QD–X–X–Cy5 18 [10]
QD–X–X–X–Cy5.5 30 [8]
QD to dye rDA (Å) 76 79 83 75
Strategy II. Metal affinity coordination Donor loss % [Eae%]
QD–Cy3 62 [67]
QD–Cy3–Cy3.5 67 92 [64]
QD–Cy3–Cy3.5–Cy5 66 96 88 [61]
QD–Cy3–Cy3.5–Cy5–Cy5.5 82 95 74 35 [25]
With missing dyes
QD–X–Cy3.5 41 [34]
QD–X–X–Cy5 14 [22]
QD–X–X–X–Cy5.5 44 [12]
QD to dye rDA (Å) 84 93 103 80

The full construct's overall estimated Eae value is 20% which is more than twice that attained in our previous QD constructs.34 Within the full construct (Fig. 4A) we do note that the QD is quenched around 80%, the Cy3 dye by 100%, and the Cy3.5 dye by >90% (Table 2) suggesting that these all function relatively well in their donor or intermediary relay capacity. The level of quenching is also slightly more than that predicted for just donor dye interactions with the next acceptor (see ESI) which again reflects the presence of multiple long range homo- and hetero FRET interactions. The Cy5 dye is, however, only quenched by <25% suggesting that it is not an efficient penultimate donor or relay. Poor Cy5 dye performance in this role has been noted previously and is suggested to arise from a variety of complex factors including local environment, free radical or dimer formation, interactions with nucleobases or neighboring photoexcited dyes, and rotational and/or conformational rigidity leading to unfavorable dipole orientation for exciton coupling.20,39,64 However, the fact that the overall Eae value is better than a summation of each dye's individual quenching suggests that some long-range interactions between the QD/Cy3/Cy3.5 and Cy5.5 help overcome the intervening Cy5's non-idealistic performance. Data from control samples with one or two intermediary/relay dyes missing support this notion. For example removing the Cy3 or Cy3.5 dye still yields Eae values of 3 and 9%, respectively, while removing the Cy5 dye drops the value from 20 to just 16% (ESI Table S1). The dendrimer's inherent ability to provide multiple overlapping and redundant pathways for exciton travel most likely contributes to the latter type of interaction.

3.5. FRET in QD-dendrimer MPWs assembled by strategy II – metal affinity coordination

The other set of constructs prepared using His6-metal affinity coordination chemistry were probed next. DNA sequences utilized here were almost identical as before, with the exception of the Cy3 modification. Due to difficulty in synthesis, Cy3 dye was not incorporated into the DNA portion of the chimeric His6-peptide-DNA strand. Instead, the complimentary strand was doubly-labeled with Cy3 at the closest possible locations as in the previous approach (Fig. 1B). In agreement with previous observations,34 after dendrimer assembly to the QD, a slight increase in QD PL was noticed (<20%, data not shown), which is attributed to better surface passivation. Fig. 5A–D depicts the spectral evolution of the QD-dendrimer constructs as consecutive acceptor dyes are added into the system along with selected controls where one or more dyes are missing in the dendrimers. A clear loss of each donor's fluorescence with concomitant acceptor sensitization was noted at each FRET step. The Cy3 and Cy3.5 dyes again appear to be transferring energy at near unity with an almost complete loss of their presence in the spectra when functioning as relays. QD donor PL also decreases as each consecutive acceptor dye was added suggesting some long-range interactions are again taking place between them. A FRET E of ∼62% was observed in the first step from the QD to Cy3 yielding an rDA of 84 Å which is 8 Å (∼10%) further than the spacing in the above covalent construct. rDA values for QD separation to the remaining Cy3.5 and Cy5 determined from control constructs where only each acceptor dye was present around the QD (Fig. 4D) place them at 93 Å and 103 Å, respectively. This is about 6–7 Å further for each placement than the above covalent configuration. The level of FRET quenching for each dye is slightly higher than what would be expected with just the next acceptor and again arises from multiple longer range homo- and heteroFRET interactions. In this case, we hypothesize that the QD's DHLA-NTA surface ligand's strong negative charge may be electrostatically influencing the dendrimer to assume a more rigid conformation away from the QD surface. The direct coordination of the His6-moiety to the QD's surface will most probably place significantly more of the dendrimeric DNA in much closer proximity to the NTA ligands than the above approach. This would, in turn, allow electrostatic repulsion to occur between the two and alter relative QD to dendrimer orientation. With the increased separation, direct QD energy transfer to each dye is somewhat mixed with a QD–Cy3 Eae value higher than the covalent construct (67 vs. 53%), that to Cy3.5 lower (34 vs. 44%), and to Cy5 higher (22 vs. 10%), see Table 2.
image file: c7me00121e-f5.tif
Fig. 5 Representative plots showing FRET evolution and progression in the QD–DNA dendrimer constructs prepared via strategy II – Hisn-metal affinity coordination chemistry. (A) PL spectra collected as sequential acceptor dyes were incorporated in the dendrimer. The black trace corresponds to the emission spectrum of QDs while others show clear signature of the FRET sensitized dye components. The QD only sample is assembled with unlabeled dendrimeric DNA. (B) Spectral evolution of control samples with one missing dye plotted superimposed over the construct with maximum dye display (red trace). A significant degree of FRET takes place despite interruption in the sequential FRET. (C) Spectra of another set of controls with two selected dyes missing. (D) Dotted traces displaying the PL spectra collected from QD-dendrimer constructs with only the one indicated acceptor dye present while the continuous traces show the intensity profiles of the directly excited dyes at the same excitation wavelength.

The QD–Cy5.5 rDA of 80 Å places it closer to the QD surface than the Cy3 dye and this result is again attributed to the same reasoning outlined above. Direct QD to Cy5.5 Eae is also more significant at 12% vs. 8% noted for the covalent construct above. The full constructs overall estimated Eae value is even higher than before achieving a remarkable 25%. As before, examination of control samples again suggests the presence of multiple long-range interactions between non-sequentially assembled dyes contributing to this high transfer efficiency. Removing the Cy3 or Cy3.5 dye yields Eae values of 11 and 20%, respectively, while removing the Cy5 dye actually increases the value to 38%, an almost 50% increase (ESI Table S2). The latter result is quite reminiscent of that seen earlier in a linear DNA photonic wire where removal of a non-performing AF610 intermediary relay dye doubled energy transfer efficiency to the next downstream acceptor.18 Cy3.5 has an excellent R0 value of 56 Å with Cy5.5 which clearly allows for high FRET E to be maintained to the latter over their separation in the dendrimer when the Cy5 is missing (this was estimated to be ∼70% which is close to the 78% value noted in experiments). Moreover, the much larger relative number of Cy5.5 acceptor dyes present, i.e. multiple acceptors per Cy3.5 donor, in conjunction with the supposition that they are located at multiple position including those that are close to the QD (and Cy3.5), would also help compensate and maintain a high level of energy transfer. Cumulatively, these results along with the poor performance of Cy5 noted above suggest that it may be functioning as more of an energy sink rather than an effective relay in the dendrimer.

4. Conclusions

In contrast to displaying individual acceptor dye labeled linear DNA photonic wires around the QDs, coupling dye-labeled dendrimers to the QDs achieves our primary goal of significantly increasing overall energy transfer efficiency through these systems. Our first QD–DNA wire systems displayed <1% efficiency and subsequent redesign and reengineering increased this to ∼10%.33,34 Replacement of the DNA wire architecture with that of a dendrimer now more than doubles this value to 20–25% depending upon the chemical attachment process utilized. This increase is the cumulative result of several contributing factors. Attaching six dendrimers each displaying two initial Cy3 dyes around each QD effectively maximizes this first FRET step by proportionally increasing the Cy3 acceptor absorption cross section.22 Indeed, FRET to the proximal Cy3 is close to the predictions expected for this system under optimal conditions, see ESI. The 2[thin space (1/6-em)]:[thin space (1/6-em)]1 dendrimer branching ratio combined with placement of 2 acceptors for each subsequent relay dye propagates this benefit through the remainder of the system. Moreover, the dendritic architecture inherently provides multiple overlapping homo- and hetero-FRET pathways for exciton travel along with contributing structural and functional redundancy to compensate for any assembly deficiencies.20 As amply demonstrated, long-range FRET steps between donor moieties and non-proximal acceptors also contribute to achieving high energy transfer flux through these systems.

Although functionally analogous to some extent, each of the QD attachment chemistries bring with them both benefits and liabilities. The covalent approach provides a robust construct that anchors the dendrimer on the QD ligand's periphery. This would still allow for other entities to be self-assembled directly to the QD surface by Hisn-Zn metal affinity in pursuit of multifunctional structures, for example, attachment of cell penetrating peptides could allow for cellular delivery of such constructs for potential biosensing.65 The issues associated with this chemistry arise from the large reagent excess required and the need for multiple rounds of purification which can adversely affect QD QY. Use of commercial amphiphilic polymer coated QDs may help to mitigate the QY effect as the QD surface is better protected in this configuration but this is at the cost of an increased QD-dye separation which can be detrimental for FRET.22,41 An alternative to this would be the use of multishell QD structures that may be more tolerant of strongly reactive chemistries at their surface.37 The EDC coupling approach can most likely be implemented with almost an ligand type on the QD surface as long as the requisite carboxyl (or cognate amine groups) are available. The second metal-affinity attachment approach is far easier and quicker to implement as it only requires mixing of the two components. The issue here is, however, one of first implementing a far more specialized chemistry to create the required chimeric His6-peptido-DNA and utilizing a QD solubilizing ligand that does not sterically hinder the DNA from attaching directly to the exposed Zn surface. Moreover, the current results suggest that the QDs surface ligands may influence the resulting DNA configuration; the latter is something that has been noted before for both DNA and peptides.66,67 In practice, the latter metal-affinity assembled QD–DNA constructs were found to be stable in solution for at least one week with refrigeration (data not shown). Similar peptide-QD conjugates were found to be stable after introduction into animal systems for more than a week as well.57,58 We note that recent synthetic approaches have focused on engineering the QD shell and increasing the thickness of the ZnS outer layer in pursuit of better quantum yields and less blinking.68 However, as elegantly shown by the Dennis Lab, this does not preclude efficient FRET nor assembly by histidine driven metal affinity.69 Although the choices are far more limited, His6-modified DNA can be coordinated to QDs functionalized with other ligands as long as they display the requisite carboxyl groups in an NTA format or capable of functionally mimicking this chelating group.41,70 Although not directly tested here per se, both approaches should allow for assembly of only one dendrimer per QD but at the cost of some unlabeled QDs in the background unless a purification step is included.65 The anomalous Cy5.5 location result is intriguing and the subject of future studies which may also serve to provide insight into the dynamics of single donor – multiple acceptor FRET configurations.

In summary, we utilize a bottom-up approach here to construct QD–DNA-dendrimer hybrid structures where QD donors and pendant dye acceptors on the dendrimer work in concert to display potent light harvesting/directing properties. In this biomimetic strategy, QDs efficiently capture light while the high dye packing density in the dendrimers facilitate subsequent propagation of the photonic energy to the redder region of the spectrum in an effective manner. The number of alternative structures available within this type of hybrid configuration are almost limitless as structural DNA maturity now allows almost any DNA scaffold architecture to be designed, assembled and attached to a QD. One intriguing hypothetical construct consists of a dendritic DNA system with multiple QDs at the periphery and which concentrates the optical energy generated by a bioluminescent protein appended to the QDs.21 This would essentially function as a hybrid chemically-driven self-illuminating nanoantenna capable of focusing its own energy to an apex. Another sensing possibility is to fuse such dendrimers with DNA that recognizes a specific target but is prebound to a quenching material such as a C60 fullerene71 or a metallic nanowire;72 interactions with the target would then free the structure and strongly increase the FRET or emission. As QD biocomposites and FRET configurations associating QDs with pendent acceptors assembled on biological scaffolds are becoming increasing popular,3,4 the hope is that the current results can provide some useful insight for designing such photonically-active emergent materials.

Conflicts of interest

There are no conflicts to declare.


The Author's gratefully acknowledge Dr. S. Diaz for assistance with data interpretation and the AIMLab at the University of Maryland for TEM assistance. Financial support is acknowledged from ONR, NRL, the NRL-NSI and LUCI grants in support of the VBFF program through the OSD.


  1. N. F. Yu, P. Genevet, M. A. Kats, F. Aieta, J. P. Tetienne, F. Capasso and Z. Gaburro, Science, 2011, 334, 333–337 CrossRef CAS PubMed.
  2. A. K. Yang, D. Q. Wang, W. J. Wang and T. W. Odom, Annu. Rev. Phys. Chem., 2017, 68, 83–99 CrossRef CAS PubMed.
  3. C. M. Spillmann and I. L. Medintz, J. Photochem. Photobiol., C, 2015, 23, 1–24 CrossRef CAS.
  4. C. E. Rowland, J. B. Delehanty, C. L. Dwyer and I. L. Medintz, Mater. Today, 2017, 20, 131–141 CrossRef CAS.
  5. B. Albinsson, J. K. Hannestad and K. Boerjesson, Coord. Chem. Rev., 2012, 256, 2399–2413 CrossRef CAS.
  6. M. F. Garcia-Parajo, J. Hernando, G. S. Mosteiro, J. P. Hoogenboom, E. van Dijk and N. F. van Hulst, ChemPhysChem, 2005, 6, 819–827 CrossRef CAS PubMed.
  7. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, New York, 2006 Search PubMed.
  8. FRET – Förster Resonance Energy Transfer From Theory to Applications, ed. I. L. Medintz and N. Hildebrandt, Wiley-VCH, Weinheim, Germany, 2013 Search PubMed.
  9. Y. J. Chen, B. Groves, R. A. Muscat and G. Seelig, Nat. Nanotechnol., 2015, 10, 748–760 CrossRef CAS PubMed.
  10. M. R. Jones, N. C. Seeman and C. A. Mirkin, Science, 2015, 347, 1260901 CrossRef PubMed.
  11. Y. R. Yang, Y. Liu and H. Yan, Bioconjugate Chem., 2015, 26, 1381–1395 CrossRef CAS PubMed.
  12. T. G. Cha, J. Pan, H. R. Chen, H. N. Robinson, X. Li, C. D. Mao and J. H. Choi, J. Am. Chem. Soc., 2015, 137, 9429–9437 CrossRef CAS PubMed.
  13. T. G. Cha, J. Pan, H. R. Chen, J. Salgado, X. Li, C. D. Mao and J. H. Choi, Nat. Nanotechnol., 2014, 9, 39–43 CrossRef CAS PubMed.
  14. J. Pan, F. R. Li, T. G. Cha, H. R. Chen and J. H. Choi, Curr. Opin. Biotechnol., 2015, 34, 56–64 CrossRef CAS PubMed.
  15. W. P. Klein, S. A. Díaz, S. Buckhout-White, J. S. Melinger, P. D. Cunningham, E. R. Goldman, M. G. Ancona, W. Kuang and I. L. Medintz, Adv. Opt. Mater., 2017 DOI:10.1002/adom.201700679.
  16. J. K. Hannestad, P. Sandin and B. Albinsson, J. Am. Chem. Soc., 2008, 130, 15889–15895 CrossRef CAS PubMed.
  17. S. A. Diaz, S. Buckhout-White, M. G. Ancona, C. M. Spillmann, E. R. Goldman, J. S. Melinger and I. L. Medintz, Adv. Opt. Mater., 2016, 4, 399–412 CrossRef CAS.
  18. C. M. Spillmann, S. Buckhout-White, E. Oh, E. R. Goldman, M. G. Ancona and I. L. Medintz, Chem. Commun., 2014, 50, 7246–7249 RSC.
  19. G. Sanchez-Mosteiro, E. van Dijk, J. Hernando, M. Heilemann, P. Tinnefeld, M. Sauer, F. Koberlin, M. Patting, M. Wahl, R. Erdmann, N. F. van Hulst and M. F. Garcia-Parajo, J. Phys. Chem. B, 2006, 110, 26349–26353 CrossRef CAS PubMed.
  20. S. Buckhout-White, C. M. Spillmann, W. R. Algar, A. Khachatrian, J. S. Melinger, E. R. Goldman, M. G. Ancona and I. L. Medintz, Nat. Commun., 2014, 5, 5615 CrossRef CAS PubMed.
  21. C. W. Brown, III, A. Samanta, S. A. Diaz, S. Buckhout-White, S. A. Walper, E. R. Goldman and I. L. Medintz, Adv. Opt. Mater., 2017, 5, 1700181 CrossRef.
  22. N. Hildebrandt, C. M. Spillmann, W. R. Algar, T. Pons, M. H. Stewart, E. Oh, K. Susumu, S. A. Diaz, J. B. Delehanty and I. L. Medintz, Chem. Rev., 2017, 117, 536–711 CrossRef CAS PubMed.
  23. W. R. Algar, H. Kim, I. L. Medintz and N. Hildebrandt, Coord. Chem. Rev., 2014, 263, 65–85 CrossRef.
  24. K. F. Chou and A. M. Dennis, Sensors, 2015, 15, 13288–13325 CrossRef PubMed.
  25. A. M. Dennis, W. J. Rhee, D. Sotto, S. N. Dublin and G. Bao, ACS Nano, 2012, 6, 2917–2924 CrossRef CAS PubMed.
  26. L. Li, F. F. Gao, J. Ye, Z. Z. Chen, Q. L. Li, W. Gao, L. F. Ji, R. R. Zhang and B. Tang, Anal. Chem., 2013, 85, 9721–9727 CrossRef CAS PubMed.
  27. W. Tang, G. C. Zhu, L. Liang and C. Y. Zhang, Analyst, 2015, 140, 5936–5943 RSC.
  28. A. Shamirian, H. S. Afsari, A. Hassan, L. W. Miller and P. T. Snee, ACS Sens., 2016, 1, 1244–1250 CrossRef CAS PubMed.
  29. A. Shamirian, H. S. Afsari, D. H. Wu, L. W. Miller and P. T. Snee, Anal. Chem., 2016, 88, 6050–6056 CrossRef CAS PubMed.
  30. A. Shamirian, A. Ghai and P. T. Snee, Sensors, 2015, 15, 13028–13051 CrossRef PubMed.
  31. S. Doughan, U. Uddayasankar, A. Peri and U. J. Krull, Anal. Chim. Acta, 2017, 962, 88–96 CrossRef CAS PubMed.
  32. U. Uddayasankar and U. J. Krull, Langmuir, 2015, 31, 8194–8204 CrossRef CAS PubMed.
  33. K. Boeneman, D. E. Prasuhn, J. B. Blanco-Canosa, P. E. Dawson, J. S. Melinger, M. Ancona, M. H. Stewart, K. Susumu, A. Huston and I. L. Medintz, J. Am. Chem. Soc., 2010, 132, 18177–18190 CrossRef CAS PubMed.
  34. C. M. Spillmann, M. G. Ancona, S. Buckhout-White, W. R. Algar, M. H. Stewart, K. Susumu, A. L. Huston, E. R. Goldman and I. L. Medintz, ACS Nano, 2013, 7, 7101–7118 CrossRef CAS PubMed.
  35. I. Medintz, Nat. Mater., 2006, 5, 842–842 CrossRef CAS PubMed.
  36. K. Susumu, E. Oh, J. B. Delehanty, J. B. Blanco-Canosa, B. J. Johnson, V. Jain, W. J. Hervey, W. R. Algar, K. Boeneman, P. E. Dawson and I. L. Medintz, J. Am. Chem. Soc., 2011, 133, 9480–9496 CrossRef CAS PubMed.
  37. K. Susumu, L. D. Field, E. Oh, M. Hunt, J. B. Delehanty, V. Paloma, P. E. Dawson, A. L. Huston and I. L. Medintz, Chem. Mater., 2017, 29, 7330–7344 CrossRef CAS.
  38. C. L. Dwyer, S. A. Diaz, S. A. Walper, A. Samanta, K. Susumu, E. Oh, S. Buckhout-White and I. L. Medintz, Chem. Mater., 2015, 27, 6490–6494 CrossRef CAS.
  39. P. D. Cunningham, A. Khachatrian, S. Buckhout-White, J. R. Deschamps, E. R. Goldman, I. L. Medintz and J. S. Melinger, J. Phys. Chem. B, 2014, 118, 14555–14565 CrossRef CAS PubMed.
  40. D. Mathur, A. Samanta, E. Oh, S. A. Diaz, K. Susumu, M. G. Ancona and I. L. Medintz, Chem. Mater., 2017, 29, 5762–5766 CrossRef CAS.
  41. J. B. Blanco-Canosa, M. Wu, K. Susumu, E. Petryayeva, T. L. Jennings, P. E. Dawson, W. R. Algar and I. L. Medintz, Coord. Chem. Rev., 2014, 263, 101–137 CrossRef.
  42. D. E. Prasuhn, A. Feltz, J. B. Blanco-Canosa, K. Susumu, M. H. Stewart, B. C. Mei, A. V. Yakovlev, C. Loukov, J. M. Mallet, M. Oheim, P. E. Dawson and I. L. Medintz, ACS Nano, 2010, 4, 5487–5497 CrossRef CAS PubMed.
  43. W. R. Algar, D. E. Prasuhn, M. H. Stewart, T. L. Jennings, J. B. Blanco-Canosa, P. E. Dawson and I. L. Medintz, Bioconjugate Chem., 2011, 22, 825–858 CrossRef CAS PubMed.
  44. G. T. Hermanson, Bioconjugate Techniques, Academic Press, San Diego, 2008 Search PubMed.
  45. M. Wu, E. Petryayeva, I. L. Medintz and W. R. Algar, Methods Mol. Biol., 2014, 1199, 215–239 CAS.
  46. Y. G. Li, Y. D. Tseng, S. Y. Kwon, L. D'Espaux, J. S. Bunch, P. L. McEuen and D. Luo, Nat. Mater., 2004, 3, 38–42 CrossRef CAS PubMed.
  47. W. C. W. Chan and S. Nie, Science, 1998, 281, 2016–2018 CrossRef CAS PubMed.
  48. H. Mattoussi, J. M. Mauro, E. R. Goldman, G. P. Anderson, V. C. Sundar, F. V. Mikulec and M. G. Bawendi, J. Am. Chem. Soc., 2000, 122, 12142–12150 CrossRef CAS.
  49. L. Berti, P. S. D'Agostino, K. Boeneman and I. L. Medintz, Nano Res., 2009, 2, 121–129 CrossRef CAS.
  50. I. L. Medintz, L. Berti, T. Pons, A. F. Grimes, D. S. English, A. Alessandrini, P. Facci and H. Mattoussi, Nano Lett., 2007, 7, 1741–1748 CrossRef CAS PubMed.
  51. D. E. Prasuhn, J. B. Blanco-Canosa, G. J. Vora, J. B. Delehanty, K. Susumu, B. C. Mei, P. E. Dawson and I. L. Medintz, ACS Nano, 2010, 4, 267–278 CrossRef CAS PubMed.
  52. S. A. Diaz, G. L. Aragones, S. Buckhout-White, X. Qiu, E. Oh, K. Susumu, J. S. Melinger, A. L. Huston, N. Hildebrandt and I. L. Medintz, J. Phys. Chem. Lett., 2017, 8, 2182–2188 CrossRef CAS PubMed.
  53. J. C. Breger, S. Buckhout-White, S. A. Walper, E. Oh, K. Susumu, M. G. Ancona and I. L. Medintz, Nano Futures, 2017, 1, 011002 CrossRef.
  54. J. H. Wang, P. J. Jiang, L. Q. Gao, Y. S. Yu, Y. Lu, L. Qiu, C. L. Wang and J. Xia, J. Nanopart. Res., 2013, 15, 1914 CrossRef.
  55. J. H. Wang, P. J. Jiang, Z. Y. Han, L. Qiu, C. L. Wang, B. Zheng and J. Xia, Langmuir, 2012, 28, 7962–7966 CrossRef CAS PubMed.
  56. J. P. Wang, Y. Y. Nie, Y. Lu, J. H. Liu, J. H. Wang, A. S. Fu, T. G. Liu and J. Xia, Langmuir, 2014, 30, 2161–2169 CrossRef CAS PubMed.
  57. K. L. Hess, E. Oh, L. H. Tostanoski, J. I. Andorko, K. Susumu, J. R. Deschamps, I. L. Medintz and C. M. Jewell, Adv. Funct. Mater., 2017, 27, 1700290 CrossRef.
  58. R. Agarwal, M. S. Domowicz, N. B. Schwartz, J. Henry, I. Medintz, J. B. Delehanty, M. H. Stewart, K. Susumu, A. L. Huston, J. R. Deschamps, P. E. Dawson, V. Palomo and G. Dawson, ACS Chem. Neurosci., 2015, 6, 494–504 CrossRef CAS PubMed.
  59. D. E. Prasuhn, J. R. Deschamps, K. Susumu, M. H. Stewart, K. Boeneman, J. B. Blanco-Canosa, P. E. Dawson and I. L. Medintz, Small, 2010, 6, 555–564 CrossRef CAS PubMed.
  60. S. K. Kailasa and S. H. Kang, Sep. Purif. Rev., 2009, 38, 242–288 CrossRef CAS.
  61. K. E. Sapsford, K. M. Tyner, B. J. Dair, J. R. Deschamps and I. L. Medintz, Anal. Chem., 2011, 83, 4453–4488 CrossRef CAS PubMed.
  62. I. L. Medintz, T. Pons, K. Susumu, K. Boeneman, A. M. Dennis, D. Farrell, J. R. Deschamps, J. S. Melinger, G. Bao and H. Mattoussi, J. Phys. Chem. C, 2009, 113, 18552–18561 CAS.
  63. E. A. Jares-Erijman and T. M. Jovin, Nat. Biotechnol., 2003, 21, 1387–1395 CrossRef CAS PubMed.
  64. C. A. M. Seidel, A. Schulz and M. H. M. Sauer, J. Phys. Chem., 1996, 100, 5541–5553 CrossRef CAS.
  65. J. C. Breger, M. Muttenthaler, J. B. Delehanty, D. A. Thompson, E. Oh, K. Susumu, J. R. Deschamps, G. P. Anderson, L. D. Field, S. A. Walper, P. E. Dawson and I. L. Medintz, Nanoscale, 2017, 9, 10447–10464 RSC.
  66. K. Boeneman, J. R. Deschamps, S. Buckhout-White, D. E. Prasuhn, J. B. Blanco-Canosa, P. E. Dawson, M. H. Stewart, K. Susumu, E. R. Goldman, M. Ancona and I. L. Medintz, ACS Nano, 2010, 4, 7253–7266 CrossRef CAS PubMed.
  67. K. B. Gemmill, S. A. Diaz, J. B. Blanco-Canosa, J. R. Deschamps, T. Pons, H. W. Liu, A. Deniz, J. Melinger, E. Oh, K. Susumu, M. H. Stewart, D. A. Hastman, S. H. North, J. B. Delehanty, P. E. Dawson and I. L. Medintz, Chem. Mater., 2015, 27, 6222–6237 CrossRef CAS.
  68. Y. Fu, K. Kim, W. Jiang, W. Yin, T. K. Ahn and H. Chae, RSC Adv., 2017, 7, 40866 RSC.
  69. M. Chern, T. T. Nguyen, A. H. Mahler and A. M. Dennis, Nanoscale, 2017, 9, 16446–16458 RSC.
  70. H. Yao, Y. Zhang, F. Xiao, Z. Xia and J. Rao, Angew. Chem., Int. Ed., 2007, 46, 4346–4349 CrossRef CAS PubMed.
  71. H. Li, Y. Zhang, Y. Luo and X. Sun, Small, 2011, 7, 1562–1568 CrossRef CAS PubMed.
  72. J. Tian, N. Cheng, Q. Liu, W. Xing and X. Sun, Angew. Chem., Int. Ed., 2015, 54, 5493–5497 CrossRef CAS PubMed.
  73. E. Oh, F. Fatemi, M. Currie, J. B. Delehanty, T. Pons, A. Fragola, S. Lévêque-Fort, R. Goswami, K. Susumu, A. L. Huston and I. L. Medintz, Part. Part. Syst. Charact., 2013, 30, 453–466 CrossRef CAS.


Electronic supplementary information (ESI) available: This includes the ligand structure, efficiency results from multiple control structures, DNA sequences, and some FRET simulations. See DOI: 10.1039/c7me00121e
Current affiliation: Ramakrishna Mission Vidyamandira, Belur Math, Howrah, WB-711202, India.

This journal is © The Royal Society of Chemistry 2018