Compact quantum dot surface modification to enable emergent behaviors in quantum dot-DNA composites

Modification of quantum dots (QDs) with single-stranded deoxyribonucleic acid (ssDNA) oligonucleotides is critical to their use in many applications, such as microarray detection,¹ Förster resonance energy transfer (FRET)-based biosensing,² and bioimaging.³ Additionally, the rise of DNA nanotechnology has enabled precise organization of inorganic nanoparticles (NPs), permitting next-generation optoelectronic and photonic devices to be realized via exploitation of NP emergent behaviors.⁴ However, the requirements for QDs conjugated to ssDNA (i.e., ssDNA-QDs) often exceed those of other biological QD conjugates, including the need for stability in solutions of wide pH range and high ionic strength. In particular, conjugation to DNA origami nanostructures can require stability in solutions with 5–20 mM Mg²⁺ concentrations.^5,6 Many applications also demand precise control over the number of ssDNAs conjugated, with some applications necessitating a single ssDNA per particle, whereas in other cases, high degrees of multivalency are desired.^7,8 For applications that focus on distance dependent emergent behaviors between QDs and other NPs, such as FRET⁹ or fluorescence quenching, all of these must be achieved while minimizing coating thickness. Thus, methods that enable precise control of ssDNA conjugation to compact QDs are needed to realize their full potential in a wide variety of fields ranging from biomedical imaging to photonics.

Because high quality QDs are primarily manufactured in the organic phase, ssDNA modification approaches necessarily require solubilization in aqueous media. Methods to obtain ssDNA-QDs can be divided into three main approaches:^10,11 (i) ssDNA incorporation in the QD crystal lattice during synthesis, (ii) direct attachment of ssDNA to the QD surface, and (iii) covalent attachment of ssDNA to the functional groups of ligands on QD surfaces. However, these methods suffer from several limitations, such as requiring a large excess of expensive ssDNA,¹² lack of control over ssDNA density,¹² thick surface coatings that limit capability to achieve emergent interactions,¹³ or loss of colloidal stability.^14,15 For example, the most common approach employed, which is also used in commercial products, consists of QD modification with a thick polymer coating that intercalates with organic ligands on the QD surface; ssDNA attachment is achieved via low yield carbodiimide chemistry routes.¹⁶ These approaches may maintain QD colloidal and photostability, but at the expense of a thick (i.e., >3 nm increase in radius) coating and low control over ssDNA density.

Several methods have been introduced to address challenges in QD conjugation, many of which rely on the high affinity of cysteine groups for the QD surface.^17,18 Recently, a cysteine-rich peptide, γ-phytochelatin-3 (PC3), was demonstrated to provide a compact coating (i.e., ∼0.8–0.9 nm increase in radius) for aqueous QDs.¹⁹ PC3 is applied through exchange with native organic ligands and presents multiple functional groups (4-COOH and 1-NH₂ per chain) available for conjugation in its final form. PC3-QDs demonstrate strong stability across a wide pH range (i.e., pH 5–10) and at high ionic strength (i.e., up to 1.5M NaCl).¹⁹ These properties are enabled by the flat conformation adopted by PC3 molecules, with multiple thiol bonds to the QD surface and preferable orientation of remaining –COOH and –NH₂ functional groups outward toward the surrounding media. However, to the best of our knowledge, PC3-QDs have only been employed in streptavidin-biotin conjugation schemes that generate large increases in composite thickness (i.e., ∼6 nm increase in radius). Whereas PC3 –COOH and –NH₂ groups are readily incorporated into low yield (∼30%) carbodiimide-based conjugation schemes,¹⁶ PC3-QDs lack functional groups required to take advantage of emerging biorthogonal conjugation methods, especially alkyne-azide “click” reactions. Click chemistry approaches provide a reliable means to achieve stoichiometric conjugation as a result of highly selective and efficient cycloaddition under mild conditions.⁸ 

Here, we combined PC3 passivation methods with copper-free alkyne-azide click chemistry to yield compact QDs conjugated to ssDNAs with precision. As model systems, we employed CdSe/ZnS QDs, most commonly used for biomedical imaging, and poly-thymine (poly-T) ssDNA sequences, which minimize nonspecific interactions with ssDNA strands and the QD surface. We also extended this work to mixed base-pair (mbp) sequences. Photophysical properties and ssDNA conjugation efficiency were quantified using absorption and fluorescence spectroscopy. Then, we demonstrated ssDNA functionality using two model systems that require compact QDs as a result of distance dependent behaviors or steric limitations. First, we explored the implementation of ssDNA-QDs as energy donors to gold nanoparticles (AuNPs) modified with complementary ssDNA, a system that demands close interactions between both particles and precise ssDNA conjugation to form QD-AuNP dimers. Next, we templated ssDNA-QDs on DNA origami hinge platforms through site-specific self-assembly, a sterically more demanding application. These data demonstrate methods for ssDNA-QD conjugation that result in a compact final product with potential for controllable conjugation yield, providing a possible platform technology for DNA-based QD devices.^6,12

CdSe/ZnS QDs (catalog Nos. CZ540-10, CZ560-10, and CZ600-10) were purchased from NN-Labs, LLC. (Fayetteville, Arizona). Gold nanoparticles (AuNPs) (catalog No. GP01-15-20) were purchased from NANOCS, LLC. (New York, NY). Phytochelatin-3 (PC3) was custom synthesized and purchased from LifeTein, LLC. (Hillsborough, New Jersey). Azide and thiol modified oligonucleotides (ssDNAs) with optional fluorophore modification for bioconjugation were custom-ordered from Integrated DNA Technologies (Coralville, IA) (see Table I for sequences). Sulfo-dibenzocyclooctyne amine (sDBCO) was purchased from Click Chemistry Tools, LLC. (Scottsdale, AZ) (catalog No. 1227).

Aqueous QD transfer was conducted using the “loops-trains-tails” approach via modification of the process described by Xu et al.¹⁹ Briefly, manufacturer-supplied QDs in toluene were transferred to chloroform. QDs in toluene (2 μM, 113 μl) were precipitated using methanol (100 μl) and redispersed in chloroform (37.5 μl). Then, pyridine (187.5 μl) was added to the freshly solvated QDs (6 μM), followed by chloroform evaporation by heating. This step exchanges some of the native octadecylamine ligands on the QD surface with pyridine. As a precaution, QDs were then centrifuged (20 817 g, 1 min) to remove any aggregates. Negligible precipitation was observed. After centrifugation, the supernatant containing suspended QDs (∼1 μM, 450 μl in pyridine) was mixed with PC3 (40 mg/ml, 100 μl in DI water), which formed a single-phase solution. Immediately, ligand exchange with PC3 was triggered by increasing solution pH above 10 by addition of tetramethylammonium hydroxide (TMAOH) (25 wt. % in MeOH, Millipore Sigma catalog No. 334901), leading to the precipitation of QDs in pyridine. The solution was quickly vortexed and then centrifuged to form a pellet. The pellet was dispersed in DI water and dialyzed overnight against 1x phosphate buffered saline (PBS) (0.1M NaH₂PO₄, 0.15M NaCl, pH 7.2, ThermoFisher Scientific, catalog No. 28372) using a 20 K molecular weight cutoff (MWCO) Slide-A-Lyzer Mini Dialysis Unit (Rockford, IL) to remove excess, unbound PC3. The dialyzed PC3-QD suspension was stored at 4°C until use.

Aqueous PC3-QDs were conjugated to oligonucleotides via click chemistry in two steps. First, PC3-QDs were conjugated to sDBCO using zero-length cross-linker carbodiimide chemistry. (Note: Carbodiimide chemistry has much higher yields for small molecules, such as the sDBCO molecules employed here, compared to ssDNA molecules most likely because of electrostatic and steric limitations. In addition, sDBCO is inexpensive and can therefore be added in large excess.) Briefly, PC3-QDs were exchanged into 2-(N-morpholino)ethanesulfonic acid (MES, ThermoFisher Scientific, catalog No. 28390) buffer (0.1M, pH 5) using a 7 K MWCO Zeba spin column (Pierce, IL) according to manufacturer’s protocol. PC3-QDs (70 μl, 2.5 μM) in MES were mixed with large excess of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC, Thermo Fisher Scientific, catalog No. 77149) and sulfo-N-hydroxysuccinimide (sulfo-NHS, Thermo Fisher Scientific, catalog No. 24525) (EDC:sulfo-NHS:QD = 50 000:50 000:1, final EDC and sulfo-NHS concentration = 100 mM) and incubated for 25 min. Next, activated QDs were exchanged into PBS (0.1M, pH 7.2) using a 40 K MWCO Zeba spin column and mixed with sulfo-DBCO (QD:sDBCO = 1:5000) and vortexed for 4 h at room temperature. Excess sDBCO was separated from sDBCO-QD conjugates using a 40 K MWCO Zeba spin column equilibrated with PBS buffer (0.1M, pH 7.2).

In the second step, azide-terminated oligonucleotides were conjugated to purified sDBCO-QDs via the strain-promoted alkyne-azide cycloaddition click reaction. sDBCO-QDs in PBS (1.5 μM QDs, determined using a Nanodrop 100 spectrophotometer) were mixed with azide-terminated oligonucleotides (QD:ssDNA = 1:30) and allowed to react overnight at 37 °C. Unconjugated oligonucleotides were separated from ssDNA-QD conjugates using 30 kDa Amicon centrifugal filters (3 washes with 1x PBS at 7000 rcf, 3 min each). QD-DNA conjugates were stored in 1x PBS buffer with 2 wt. % dissolved polyethylene glycol 20 kDa (PEG-20 k, Millipore Sigma, Cat. No. 95172) at 4 °C until use.

AuNP quenchers were conjugated to thiol-terminated oligonucleotides using an instantaneous, low pH-assisted conjugation protocol.²⁰ Thiol-modified oligonucleotides received in stable disulfide forms (100 μM, 50 μl) were first reduced by the reaction with 20 μl of freshly prepared 0.5M DTT solution in sodium phosphate buffer (0.5M, pH 8.4) for 1 h at room temperature. Reduced oligos were purified from DTT using a size exclusion chromatography column (GE NAP-10) according to the manufacturer’s protocol. The concentration of reduced oligos was determined by UV-vis spectroscopy by comparing their absorbance at 260 nm (ssDNA peak) with a standard curve. Then, AuNPs were mixed with reduced oligos (AuNP:oligo = 1:225) for 1 min at room temperature. Next, citrate buffer (500 mM, pH 3) was added to a final citrate buffer concentration of 10 mM and allowed to react for 3 min. The solution was then purified from unconjugated, excess ssDNA and exchanged into 1x PBS buffer using a 100 kDa Amicon centrifugal filter. The ssDNA-AuNP conjugates were stored in 1x PBS at 4 °C until use.

QDs and AuNPs presenting complementary ssDNA (Table I) were mixed in desired ratios in PBS with 2 wt. % PEG-20k at a fixed QD concentration of 3 nM and allowed to self-assemble overnight at 4 °C. For control experiments, QDs and AuNPs presenting noncomplementary ssDNA were mixed following the same method. Both the samples were prepared using green QDs (i.e., CZ540) and 15 nm AuNPs at QD:AuNP molar ratios of 2:1, a final QD concentration of 3 nM, and a PEG concentration of 2% w/v in physiological 0.1M phosphate buffer saline (PBS) at pH 7. Samples were stored overnight before fluorescence measurement.

Briefly, DNA origami hinges were assembled in a solution of 20 nM m13mp18 bacteriophage scaffold DNA, 200 nM of each staple strand, 1 mM EDTA (ethylenediaminetetraacetic acid), 5 mM NaCl (sodium chloride), 5 mM Tris (tris(hydroxymethyl)aminomethane), and 18 mM MgCl₂ (magnesium chloride) as previously described.²¹ ssDNA-QDs dissolved in 12.5 mM MgCl₂ were mixed with DNA origami hinges presenting complementary ssDNA on both arms at QD:Hinge ratios of 5:1 and incubating at 55 °C for 10 min before cooling to room temperature. QD-DNA origami hinge conjugates were purified by gel electrophoresis {i.e., 2% agarose, 0.5X [Tris, Boric acid, and Ethylenediaminetetraacetic acid (EDTA), TBE], 11 mM MgCl₂} before transmission electron microscopy (TEM) imaging.

Absorbance measurements were used to confirm conjugation and determine NP concentrations. Absorption spectra were collected using a Genesys 6 UV-Visible or Nanodrop 1000 spectrophotometer.

Fluorescence analysis using a PTI QuantaMaster fluorometer was performed to calculate quantum yields (QY), monitor QD bioconjugation to dye-labeled molecules, and analyze fluorescence quenching between QDs and AuNPs. Quantum yield was calculated using the equation described in Ref. 22,

where Φ_x, F_x, f_x(λ_ex), and η_x are the quantum yield, spectrally integrated fluorescence flux, absorption factor, and refractive index of the medium, respectively, and subscripts QS and St refer to the QD sample and a standard (Rhodamine 6G, ThermoFisher Scientific, catalog No. 252433), respectively.

For experiments optimizing conjugation, NH₂-terminated dye molecules were employed and successful conjugation was monitored by measuring the fluorescence spectra of samples after purification. The presence of both QD and dye fluorescence peaks indicated successful conjugate formation. Efficiency of ssDNA conjugation was monitored by employing dye-labeled ssDNA such that the oligonucleotide sequence was sandwiched between the –NH₂ group and the dye molecule (i.e., NH₂-ssDNA-dye). The presence of both QD and dye fluorescence peaks in purified samples indicated successful reaction. For determining the number of ssDNA conjugated per QD, dye-tagged ssDNA was also used. The intensity of QD and dye-ssDNA fluorescence peaks was compared to their respective standard curves to calculate the ratio of ssDNA per QD. All measurements were conducted at concentrations that provided the measurable signal for both QDs and conjugated ssDNA-dye molecules (∼150 nM).

Fluorescence quenching between QD and AuNP pairs was analyzed by comparing the fluorescence of hybridized QD-AuNP complexes (complementary sequence) with the fluorescence of unhybridized QD/AuNP solutions (noncomplementary sequence) at the same concentration in 2 wt. % PEG-20 K at a QD:AuNP ratio of 2:1 and at 3 nM ssDNA-QD. PEG-20 K molecules were included to enable depletion stabilization of complexes.

QDs, AuNPs, and corresponding conjugates were imaged using an FEI Tecnai G2 Bio Twin TEM. Aqueous samples were deposited on ultrathin carbon coated copper grids, 300 mesh (Ted Pella, Redding, CA). For imaging QD-DNA origami conjugates, TEM grids were negatively stained with 2% uranyl formate as previously described.²³ 

Gel electrophoresis was performed to validate ssDNA conjugation and to determine its efficiency using Cy5 dye-tagged ssDNA. During gel electrophoresis, unpurified ssDNA-QD samples (100 μM, 15 μl), ssDNA, and PC3-QDs were analyzed using a 2 wt. % agarose gel prepared in TE buffer at 110 V for 45 min. After completion, QD and dye-tagged ssDNA fluorescence was assessed using an Imager-Typhoon Trio (GE Healthcare) scanner with the Cy2-FRET and Cy5 excitation settings, respectively. Gel images were analyzed using NIH ImageJ to determine QD and ssDNA locations in the gel. The gel analysis feature of ImageJ was used to measure the ratio of fluorescence from conjugated vs unconjugated ssDNA to calculate the conjugation efficiency (conj%). The amount of ssDNA per QD was then calculated by

The highest quality QDs are manufactured in the organic phase; thus, they must first be transferred into the aqueous phase prior to ssDNA conjugation. Here, PC3 ligands were used to solubilize QDs because they exhibit numerous sulfur groups that form strong multidentate linkages to the QD surface and outward-facing –COOH and –NH₂ functional groups that increase the colloidal stability through electrostatic repulsion and provide sites for ssDNA attachment.¹⁹ Successful PC3 coating and QD phase transfer were performed using a modified version of the method described by Xu et al.¹⁹ in which QDs were not precipitated before dissolution in pyridine, reducing QDs losses resulting from this step (see the supplementary material). With this modified approach, nearly 100% phase-transfer yield was observed.

In contrast to prior reports,¹⁹ PC3-QDs exhibited a small red-shift in peak emission wavelength (λ_em) (approximately few nanometers) after the transfer process (Fig. 1, Table II), which was more pronounced for smaller QDs than larger ones (i.e., λ₅₄₀ > λ₅₆₀ > λ₆₀₀). Size-dependent differences most likely reflect the reduced stability of smaller QDs, which have greater surface to volume ratio. Differences with respect to prior reports likely result from the native surface ligands employed: octadecylamine (ODA) in the current study vs trioctylphosphine oxide (TOPO).¹⁹ Primary amine ligands, such as ODA, exert compressive stress on the QD lattice in contrast to TOPO, which exerts a tensile stress.²⁴ The compressive stress of ODA is also responsible for the high QY of ODA-coated QDs. As a result, ligand exchange of ODA with PC3 may be accompanied by stretching the QD lattice, effectively increasing the size of PC3-QDs and yielding a red-shift in emission wavelength. As is typical of aqueous phase transfer processes, QY was also reduced from that observed for organic QDs (Table II) and was slightly lower than that reported for TOPO-capped QDs transferred via PC3 (i.e., 47%), most likely as a result of released lattice strain.

Initially, we attempted to conjugate ssDNA molecules to PC3-QDs using carbodiimide chemistry as previously described for streptavidin conjugation;¹⁹ however, we were unsuccessful (see Fig. 1 and Table 1 of the supplementary material). Thus, we adopted a click chemistry approach. However, PC3-QDs inherently lack alkyne or azide groups necessary for this reaction. PC3-QDs display two functionalities: –NH₂ and –COOH groups that could be used to add “click”-able groups via homo-bifunctional NHS-ester or carbodiimide zero-length cross-linker chemistries, respectively.²⁵ However, the wide availability of azide-terminated oligonucleotides and greater number of –COOH groups on PC3-QDs strongly favor their use for alkyne modification.

Thus, PC3 –COOH groups were first modified with sDBCO using carbodiimide chemistry (Fig. 2), which is much more efficient for small molecules than the ssDNAs previously attempted. QD colloidal stability was enhanced by use of sDBCO instead of DBCO, and PEG -20k enabled depletion stabilization (see Fig. 2 of the supplementary material). This approach overcomes limitations of carbodiimide conjugation processes by enabling rigid and selective conjugation pathways in the latter part of the nested conjugation. Bio-orthogonal chemistry, specifically, click chemistry, proceeds via an alkyne-azide cycloaddition reaction that has been demonstrated to be highly reproducible and efficient under mild reaction conditions.⁸ Furthermore, unlike the carbodiimide chemistry, click chemistry enables the specific reaction between alkyne-azide groups without side-reactions, such as hydrolysis seen in carbodiimide alternatives.²⁶ 

To generate ssDNA-QD conjugates, sDBCO-QDs were reacted with N₃-terminated ssDNA. Reaction progress was monitored using ssDNAs modified with both N₃ and Cy5 fluorophores (Table I). Fluorescence peaks for both QDs and Cy5 dye were observed in purified conjugates, indicating the successful reaction [Fig. 3(a)], which was further confirmed by gel electrophoresis [Figs. 3(b) and 3(c)]. [Note: The mobility difference of free Cy5-ssDNA in ssDNA and ssDNA-QD (U) lanes likely results from differences in buffer conditions during gel electrophoresis, as has previously been observed for ssDNA strands.²⁷] Fluorescence and gel electrophoresis analysis indicated a density of ∼0.5 ssDNA per QD using this approach; thus, some QDs were devoid of ssDNA. This may result from inconsistencies inherent in the “shotgun” carbodiimide chemistry employed in sDBCO attachment. With a target of ∼1 ssDNA per QD desired to achieve dimerization with AuNPs and avoid uncontrolled assembly, this was a satisfactory result. This procedure was used to reproducibly conjugate ssDNA to QDs of different sizes (i.e., λ₅₄₀ or λ₆₀₀) and with ssDNAs of different lengths (i.e., 10, 12, and 16 bp) and sequences (i.e., mbp or poly-T), indicating its robustness (Figs. 3 and 4 of the supplementary material).

To evaluate the effectiveness of ssDNA-QDs, we next evaluated them in a fluorescence quenching application. In close-proximity to AuNPs, near-field effects emerge that modify QD fluorescence through static or dynamic charge or energy transfer,²⁸ FRET,²⁹ surface field enhancement,³⁰ or nanometal surface energy transfer (NSET).^4,9 Near-field effects are extremely sensitive to interparticle spacing and decay as a function of 1/sⁿ (n ≥ 1). DNA offers tremendous potential to finely, reversibly, and reproducibly tune QD devices for these applications. Furthermore, unlike many coatings used for QD aqueous solubilization,^13,31 the compact nature of the PC3 coating enables QDs to interface more intimately with the surrounding environment (i.e., within ∼1 nm) while maintaining stability in physiological and high-salt conditions.

Generally, changes in colloidal QD fluorescence in the presence of AuNPs can be broadly divided into two categories: (i) those arising from the inner-filter effect (IFE) and (ii) those arising from near-field effects.²⁸ The IFE is an apparent decrease in fluorescence intensity resulting from attenuation of the excitation or emission beam at high concentration or in the presence of a highly absorbing material (e.g., AuNPs). Thus, the IFE effect results in a decrease in fluorescence intensity that is not a true quenching response and is not indicative of the formation of DNA-hybridized NP pairs. The IFE can be empirically eliminated by decreasing sample OD < 0.05;³² however, this was not possible here because decreasing the OD of the sample containing AuNPs < 0.05 resulted in QD fluorescence below the spectrophotometer detection limit. Alternatively, steady-state QD fluorescence can be monitored in the presence of AuNPs initially as complexes form.^4,9 ssDNA-QDs (T₁₆-QD) were incubated with noncomplementary ssDNA-AuNPs (T₁₄-AuNPs), which should not form hybridized pairs (NC) and complementary ssDNA-AuNPs (A₂₀-AuNPs) (C). Their fluorescence was compared (Fig. 4), and the % decline of the entire sample was calculated as

where F_AuNP−c and F_AuNP−nc are the steady-state peak fluorescence intensity of hybridized QD-AuNP composites and the noncomplementary ssDNA-QD and -AuNP mixture, respectively.

An ∼75% decline in fluorescence was observed for hybridized complementary samples compared to mixtures of QDs and AuNPs presenting noncomplimentary ssDNA (Fig. 4). However, this observation alone does not necessarily correlate with the quenching per hybridized NP pair, as the solution likely consists of a mixture of free and modified QDs based on fluorescence and electrophoresis results of ssDNA/QD. Given these factors, the change in QD fluorescence attributable to hybridization with AuNPs (S) can be represented as follows:²⁸ 

where F_c and F_nc are the measured fluorescence of samples incubated with C and NC AuNPs, respectively, and χ_AuNP is the fraction of AuNP-hybridized QDs.

The latter was determined experimentally by centrifuging C and NC solutions and measuring supernatant fluorescence after removing the large, high-density AuNPs and their conjugates. χ_AuNP is then equal to $Fncsuper−FcsuperFncsuper$⁠. The first term of Eq. (4) denotes the contribution of free, nonhybridized QDs to the fluorescence signal, whereas the second term denotes the contribution of hybridized composites. Based on this analysis (Table III), ∼84% of the QDs in the solution formed hybridized AuNP-QD pairs, resulting in a near-field effect of 0.11 and an effective quenching of ∼89.3% per QD, calculated as 100(1 − S).

This analysis suggests that near field effects are the dominant cause of fluorescence loss in C solutions. Furthermore, these results most likely originate from energy transfer between NPs. It has previously been demonstrated that static or dynamic charge or energy transfer, surface field enhancement, and NSET are negligible at low concentrations,²⁸ for AuNPs ≤ 30 nm in diameter,³³ and for small polarizability differences for fluorophore-quencher pairs,⁴ respectively. Thus, the most likely explanation for reduced fluorescence intensity via near field mechanisms is enhanced fluorescence quenching resulting from the significant spectral overlap between the emission of green QDs and AuNP absorbance and the reduced separation distance achieved by DNA hybridization.³⁴ At the QD:AuNP ratio (2:1) employed, quenching efficiency was significantly greater than that reported in studies using organic dye quenchers (i.e., <30% at donor:acceptor ratios of 1:1) because of the high absorption cross section of AuNPs.^35,36 Comparable quenching efficiencies can still be achieved using organic quenchers by placing them very close to the QD surface;³⁷ however, such systems suffer from poor performance because of photobleaching. Furthermore, the quenching efficiency observed here was also greater than that reported for many previous QD-AuNP systems,³⁸ most likely as a result of reduced inter-NP spacing enabled by DNA hybridization vs the use of large biomolecules (i.e., proteases) or because of low spectral overlap in those systems. Furthermore, the current system displayed comparable or even slightly higher quenching efficiency than QD-ssDNA-AuNP systems with similar spectral overlap, despite increased ssDNA chain length [i.e., 16 base pairs (bp) in the current system vs 7–10 bp in prior reports^9,39]. This difference can be attributed to the thinner coating provided by PC3 on the QD surface that reduces inter-NP spacing.

Nevertheless, quenching efficiency was lower than theoretical; ∼99% efficiency would be expected for the QDs and AuNPs used in this system [i.e., see calculations in the supplementary material for details: absorbance spectral overlap = 1.34 × 10¹⁹ M⁻¹ cm⁻¹ nm⁴, interparticle spacing(s) = 6.7 nm (calculated based on the DNA chain length), and Förster radius (R_o) = 16.6 nm]. This may be explained by quenching efficiency dependence on donor valency (n), i.e., number of donors per quencher,⁴⁰ 

where S_FRET and R_o are the quenching efficiency and Förster radius, respectively.

The current system presents variable valency (i.e., some QDs do not contain ssDNA) because of differences in ssDNA number per QD. Although most QDs formed hybridized pairs with AuNPs, which were in excess, 16% failed to do so. Improved control over ssDNA conjugation through further optimization of click-based approaches would improve these results. Similarly, the centrifugal filtration purification approach employed here only separates free ssDNA from conjugates, not unmodified QDs, which we confirm to be present in solution based on fluorescence and electrophoresis results. Thus, these results could be improved by alternative separation schemes that enhance purification. Nonetheless, these data show that compact coating methods, such as those provided by PC3, combined with click chemistry-enabled ssDNA modification provide a robust approach for QD energy transfer materials.

Next, we examined the ability of ssDNA QDs to be integrated with DNA origami materials. DNA origami can serve as a scaffold to precisely assemble NPs into complex 3D arrangements for potential applications in energy and photonics.⁴¹ However, most DNA origami studies using NPs have been limited to AuNPs,⁴¹ most likely because of their ease of ssDNA modification and stability at high ionic strength conditions (i.e., 5–20 mM divalent Mg) required to stabilize origamis.⁵ Few reports demonstrate QD attachment to DNA origami,^12,42 despite their obvious potential in optoelectrical devices. Some of these studies employ thick streptavidin coatings for composite formation that strongly limit distance-dependent emergent interactions between NPs and fail to provide the highly specific programmability achievable via DNA hybridization.⁴² Thus, as a second model system, we evaluated the ability of compact ssDNA-QDs to hybridize with sterically complex DNA origami platforms.

In these experiments, ssDNA-QDs were incubated with DNA origami hinges²¹ that have previously been used for AuNP templating⁴³ with binding sites (complementary overhangs) on the top and bottom arms [Fig. 5(a)]. QDs integrated with these platforms, as indicated by their position between hinge vertex arms in TEM images [Fig. 5(b)]. In most cases, QD binding resulted in hinges changing to a closed configuration, indicating a single QD is bound to both hinge arms. Although in some cases, QDs were bound to only one hinge arm. This may result from only 1 ssDNA/QD, which would enable only one complementary DNA pair to form. These data are similar to reports using the DNA embedding approach for QD templating on DNA origami platforms;^4,12 however, our method does not require a large excess of ssDNA for successful conjugation because of the specificity of click chemistry reactions. Combined with compact PC3 coatings that augment NP emergent interactions through reduced separation distances, this approach may provide a more feasible, less expensive route to macroscale materials with energy applications.

Here, we present a novel method for conjugation of ssDNA to QDs that combines a compact, multivalent coating methodology with mild, strain-promoted alkyne-azide cycloaddition click reactions to yield stable conjugates at high yield. This approach enhances potential for emergent interactions between QDs by reducing the thickness of coatings required to transfer QDs to the aqueous phase, while minimizing loss of fluorescence and colloidal stability. Additionally, this method provides an alternative to popular carbodiimide chemistry approaches that are inefficient and nonspecific, particularly when large, monovalent, charged molecules (e.g., ssDNA) are employed. Successful conjugation was confirmed by evaluating ssDNA hybridization in two platforms: fluorescence quenching with AuNP conjugates and conjugation to DNA origami materials. The resulting compact ssDNA-QDs hybridized with complementary ssDNA-AuNPs (2:1 QD:AuNP ratio), resulting in 84% hybridized pairs and a quenching efficiency of 89% per hybridized pair. Furthermore, enhanced colloidal stability of ssDNA-QDs enabled ready incorporation with DNA origami templating platforms despite the requirement of high ionic strength buffers to stabilize these systems. Improved control of conjugation reactions and subsequent separation steps would enhance these methods, enabling modification with specified numbers of DNA strands for the highly controlled reaction. These results demonstrate ssDNA-PC3-QDs generated using click chemistry approaches as strong candidates for applications that leverage emergent NP interactions, such as FRET and optoelectronic devices.

See the supplementary material for additional methods for PC3 phase transfer and carbodiimide conjugation of ssDNA, Figs. 1–4 and Table 1.

The authors gratefully acknowledge funding provided by the National Science Foundation (NSF), under Award No. DBI-1555470, and the U.S. Department of Energy (DOE), Office of Science, Basic Energy Sciences (BES), under Award No. DE-SC0017270. DNA modification chemistry and FRET studies were funded by the NSF, whereas DNA origami studies were funded by the DOE. In accordance with ethical obligation as a researcher, J.O.W. reports that she has financial and business interest in a company (i.e., Core Quantum Technologies) that may be affected by the research reported in the enclosed paper. J.O.W. has disclosed that interest fully to the publishers and has in place an approved plan for managing any potential conflicts arising from that involvement. This work was not funded by Core Quantum Technologies, and the opinions represented are those of the authors.