Energy transfer assays using quantum dot – gold nanoparticle complexes:

Optimizing oligonucleotide assay configuration using monovalently conjugated

quantum dots

Uvaraj Uddayasankar and Ulrich J Krull

Abstract

The energy transfer between quantum dots (QDs) and gold nanoparticles (AuNPs)

represents a popular transduction scheme in analytical assays that use nanomaterials. The

impact of the spatial arrangement of the two types of nanoparticles on analytical

performance has now been evaluated using a nucleic acid strand displacement assay. The

first spatial arrangement, Configuration 1, represents the assembly of a number of

monovalently functionalized QD-oligonucleotide conjugates around a single central

AuNP that was functionalized with complementary oligonucleotide sequences. The

assembly of these complexes, and subsequent disassembly via target oligonucleotide

mediated displacement, were used to evaluate energy transfer efficiencies. Furthermore,

the inner filter effect of AuNPs on the fluorescence intensity of the QD was studied.

AuNPs of three different diameters (6 nm, 13 nm and 30 nm) were used in these studies.

Configuration 2 was based on the placement of monovalently functionalized AuNP-

oligonucleotide conjugates around a single QD that was functionalized with a

complementary oligonucleotide. The optimal assay configuration, established by

evaluating energy transfer efficiencies and inner filter effects, was obtained by arranging

at most 15 QDs around the 13 nm AuNP (Configuration 1). These assays provided a 2.5

fold change in fluorescence intensity in the presence of target oligonucleotides. To obtain

the same response with Configuration 2 required the placement of three 6 nm AuNPs

around the QD. But the inner filter effect resulted in Configuration 2 having a 5-fold

lower fluorescence intensity when compared to Configuration 1. The use of low-cost

detection systems (digital camera) further emphasized the higher analytical performance

of Configuration 1. Response curves obtained using these detection systems

demonstrated that Configuration 1 has better sensitivities (10-fold higher) and limits of

detection (2-fold lower) when compared to Configuration 2. This study provides an

important framework for the development of sensitive assays using gold nanoparticles

and quantum dots.

Introduction

Assays based on energy transfer using quantum dots (QDs) as donors, and

metallic nanoparticles such as gold nanoparticles (AuNPs) as acceptors, have garnered

significant attention in recent years{Algar:2014cb}. While QDs have many favorable

optical properties, their physical size limits energy transfer efficiencies, {Clapp:2004kc}

thus requiring efficient energy acceptors. Metallic nanoparticles provide for competitive

energy transfer efficiencies due to their high extinction coefficients across a broad range

of wavelengths and the availability of a 2-D surface to which energy is transferred

{Singh:2010us}. While the QD-AuNP system has been utilized for the detection of

analytes such as nucleic acids {Dyadyusha:2005bp, Yeh:2010jo}, proteases

{Choi:2012ja, Kim:2008id} and inorganic ions {Li:2011ir}, the widespread adoption of

this system requires a better understanding of factors that influence its analytical

performance. Two of these factors include the inner filter effect of AuNPs and the

relative spatial arrangement of the two types of nanoparticles.

The inner filter effect refers to the attenuation of excitation source intensity and

emitted radiation due to the high extinction coefficient of AuNPs at the respective

wavelengths. Recent studies have highlighted the importance of taking this effect into

account when investigating energy transfer processes. {Han:2012iy, Zhang:2015fu}

Furthermore, it would influence the analytical performance of assays by limiting the

amount of light that may be detected. To study the impact of inner filter effect on assay

performance requires the consideration of two factors. AuNP extinction coefficient is a

function of its size, with larger AuNPs having a greater extinction coefficient.

{Haiss:2007co} However, larger AuNPs may also accommodate a greater number of

QDs on its surface. Thus, an optimal configuration may exist where the AuNP is large

enough to accommodate a sufficient number of QDs to minimize the impact of the inner

filter effect. To ensure optimal assay design, it is vital to understand and control the inner

filter effect.

Complexes of QDs and AuNPs can be created using complementary

oligonucleotide conjugates for formation of desired structures by virtue of hybridization.

{Nie:2009bp}But the large surface areas of both QDs and AuNPs necessitate stringent

control of the functionalization of these nanomaterials. The high surface area of

nanomaterials often leads to a distribution of ligand arrangements on the nanoparticle

surface. {Pons:2006kr} This increases the permutations in which the two nanoparticles

may come together, negatively impacting the performance of the assay by the reality that

the dispersity of structures provides for a dispersity of quantitative responses. Formation

of aggregates is also unavoidable as a result of the multivalency of the functionalized

nanoparticles. Due to these complications, investigations using large gold nanoparticles

have been limited. Most early studies have focused on the use of small (≤ 5 nm) AuNPs

prepared stoichiometrically to achieve monovalent conjugation (Figure 1, Configuration

2) {Pons:2007do, Choi:2012ja, Kim:2008id}. However, the conjugation of multiple

AuNPs per QD is fraught with challenges as there exists a statistical distribution of

AuNP-QD conjugate ratios, especially at low equivalences. This distribution is known to

impact energy transfer efficiencies. {Beane:2014cf} The use of monovalently

functionalized QDs may overcome some of these challenges.

Figure 1 (a) Schematic illustration of the assay configurations investigated. Configuration 1 arranges multiple QDs around a single AuNP while configuration 2 places multiple AuNPs around a single QD. (b) Toehold mediated strand displacement assay.

An alternative arrangement involves implementation of QDs that are

monovalently functionalized with an oligonucleotide. These monovalently conjugated

QDs are assembled on AuNPs (Configuration 1; Figure 1a). In this manuscript, the

analytical performance of multiple QDs surrounding a AuNP is compared to that of the

usual configuration where a single QD is surrounded by multiple small AuNPs

(Configuration 2; Figure 1a). For the two configurations, inner filter effects and the

quenching efficiencies were evaluated to establish optimal performance criteria. A toe-

hold mediated DNA displacement assay served as a representative assay design to assess

analytical performance (Figure 1b). In this assay, the AuNPs and QDs were each

functionalized with a different DNA strand (probe and displacer). The complementarity

of the sequences led to the formation of AuNP-QD complexes via DNA hybridization.

Introduction of a target DNA strand caused the disassembly of the complex due to the

higher stability of the probe-target hybrid (compared to the probe-displacer hybrid).

Using AuNPs of three different sizes (diameter of 6, 13 and 30 nm), the influence of the

inner filter effect and energy transfer processes were individually investigated for

Configuration 1 and Configuration 2. The ability to assemble and disassemble the QD-

AuNP conjugates (via the displacement assay) enabled the measurement of QD

fluorescence intensity when bound to and also when separated from the AuNPs. The

differences in fluorescence intensity provided insight about the energy transfer process.

For independent investigations of the inner filter effect, QDs and AuNPs that did not

interact with each other were prepared and their fluorescence intensities were monitored

at varying ratios of QD to AuNP. Finally, the response of the various configurations to

changing target concentrations was measured to understand how to optimize analytical

performance.

Experimental

A detailed list of materials, instrumentation and methods is available in the

Supporting Information.

Inner filter effect

Investigations of the inner filter effect required nanoparticles without any DNA,

but protected with mPEG ligands to mitigate non-specific interactions. All QDs were

coated with a DHLA mPEG (MW 750 g mol-1) ligand, while the AuNPs were coated with

thiol-functionalized mPEG (MW 800 g mol-1). Details of the preparation of these

nanoparticles can be found in the Supporting Information

For Configuration 1, the QDs were kept at a constant concentration of 50 nM,

with the amount of AuNPs varied to obtain the required ratios of QD to AuNP. A

solution of QDs at a concentration of 50 nM, without any AuNPs, served as a control.

For Configuration 2, the AuNP (diameter: 6 nm) was kept at a constant

concentration of 50 nM, with the amount of QDs varied to obtain the required ratios of

AuNP to QD. For the various concentrations of QDs used, a similar solution was

prepared without any AuNPs to serve as the control sample.

The luminescence of the prepared solutions were determined with a PTI

QuantaMaster spectrofluorimeter. The excitation wavelength was 405 nm, with the

fluorescence emission collected at 480 – 580 nm for QD525, 520 – 620 nm for QD575

and 580 – 680 nm for QD630. The fraction of retained fluorescence was calculated as:

Fraction of initial fluorescence intensity=I AuNP

I No AuNP 1

where IAuNP refers to the fluorescence intensity of a particular concentration of QDs in the

presence of gold nanoparticles, and INo AuNP refers to the fluorescence intensity of a

solution with the same QD concentration but without any gold nanoparticles. The QD

emission peaks were fit to a Gaussian function and the area of the Gaussian curve

represented the fluorescence intensities.

Assay assembly

Details on the preparation of the various nanoparticle – DNA conjugates is

available in the Supporting Information

Configuration 1:

QD-DNA monoconjugates ([QD] = 100 nM) were incubated with varying

amounts of DNA functionalized AuNPs in borate buffer supplemented with sodium

chloride (final concentration: 0.5 M). In addition to the QD-DNA conjugates, a quantity

of QD630 equal to QD-DNA monoconjugate was also added to the sample to serve as an

internal standard. The solutions were incubated at room temperature for 1 hour to ensure

complete hybridization. Prior to measurements, the samples were diluted such that the

QD concentration was 50 nM.

Configuration 2:

AuNP-DNA monoconjugates ([AuNP] = 50 nM) were incubated in borate buffer

that was supplemented with sodium chloride (0.5 M) with varying amounts of QDs that

had been functionalized with multiple DNA strands. The solution was spiked with a

quantity of QD630 at a concentration equal to that of the QD-DNA conjugates to serve as

an internal standard. The solutions were incubated at room temperature for 1 hour to

ensure complete hybridization. Prior to measurements, the samples were diluted such that

the AuNP concentration was 50 nM.

Fluorescence intensity measurements

The fluorescence intensities of all solutions were measured using an

epifluorescence microscope (described in the Supporting Information). The

epifluorescence microscope made it possible to collect emission data from small solution

volumes (≤10 μL). The fluorescence spectrum for the different samples were obtained,

and then normalized to the peak intensity corresponding to QD630. The normalized

QD525 spectrum was fit to a Gaussian function to enable calculation of peak area.

Response to target hybridization was obtained by measuring the fluorescence

intensity of solutions where the QD-AuNP complexes were incubated with the specified

amount of target for at least 1 hour. Care was taken to ensure that the monovalently

functionalized nanoparticles were always at 50 nM prior to measuring the fluorescence

intensity.

The contrast ratio was obtained as per the following equation,

Contrast ratio=I TGT

I 02

where ITGT corresponds to the area of the emission envelop for a solution with target, and

I0 refers to the area of the emission spectrum for a solution without any target.

For most measurements, the contrast ratios were measured by incubating the QD-

AuNP complexes with an excess of target (>100 equiv.) for at least 1 hour. For the

calibration curve measurements, the QD-AuNP complexes were incubated with a series

of target amounts for at least 1 hour.

Fluorescence intensity determination from an iPad camera

The plastic tubes containing QD-AuNP conjugates that were incubated with

varying amounts of target were placed at a distance of 5 cm from a UV lamp (365 nm;

UVGL-58, LW/SW, 6W; The Science Company, Denver, CO, USA), with the lamp

coplanar to the tubes. The fluorescent image was collected 20 cm above the samples (90o

with respect to the light source) using a camera that was integral to an Apple iPad 4

(Cupertino, CA). The distance of the camera from the samples was chosen to ensure all

tubes were captured in a single image frame. Measurements were taken in a dark

environment to reduce background due to ambient light. The images were obtained under

the default settings without any special precautions.

The image was processed using ImageJ, where the RGB color splitting function

was used to generate the images of the green channel (QD525) and the red channel

(QD630). The intensities of the green channel were then normalized to that of the red

channel, and this served as the response signal for the assay. The contrast ratios were

calculated as described above.

Results and discussion

Assay design

A toe-hold mediated DNA strand displacement assay served as a model system to

optimize assay configurations that utilized quantum dots (QDs) as energy donors and

gold nanoparticles (AuNPs) as acceptors. The displacement assay required initial

formation of a DNA hybrid between a probe strand and a shorter displacer strand. Each

strand was immobilized onto a different nanoparticle, and upon DNA hybridization the

nanoparticles were brought in close proximity to facilitate efficient energy transfer. For

this system of nanoparticles, the displacer strand had to be at least 16 bases long

(Supporting Information, Figure S3). Shorter displacer strands resulted in lower

hybridization efficiencies (higher background), while longer DNA strands did not lead to

further improvements in hybridization efficiency. The target DNA was complementary to

the probe strand and was longer than the displacer strand. The probe-target hybrid had 30

complementary base pairs, compared to a maximum of 16-20 base pairs for the probe-

displacer hybrid. This ensured that the probe-target hybrid was more thermodynamically

stable than the probe-displacer hybrid, favoring its formation in the presence of target

DNA strands. Hybridization of the target DNA strand with the probe strand caused the

disassembly of the AuNP-QD complex, thus decreasing the energy transfer efficiency.

Two different configurations were investigated in these experiments (Figure 1).

Configuration 1 refers to an assay in which QDs were arranged around a single AuNP by

hybridization of complementary DNA strands. Configuration 2 made use of a single QD

that was surrounded by multiple AuNPs. For both configurations, the DNA sequences

were designed to ensure minimal separation distance between the two nanoparticles. This

required the placement of both nanoparticles (QDs and AuNPs) on the same end of the

DNA hybrid. Depending on the configuration, monovalent functionalization of the QD

(configuration 1) or the AuNP (configuration 2) with the probe DNA strand was required.

The monovalent functionalized nanoparticles were selected to provide for reduced steric

and electrostatic hindrance towards the binding of long target DNA strands. Monovalent

conjugates also prevented the formation of crosslinked aggregates.

The QDs were functionalized with a bidentate poly(ethylene glycol) methyl ether

ligand (DHLA – mPEG), which is known to provide QDs with excellent colloidal

stability under a variety of solution conditions (pH and ionic strength), and also to

prevent non-specific interactions with biomolecules and other interfaces.

{Susumu:2007fy} DNA conjugation onto the QDs was achieved using the well-known

interactions of a hexahistidine moiety with the inorganic surface of the QDs.

{Medintz:2007ch} AuNPs were functionalized with thiol terminated DNA strands using

the “salt aging” protocol discussed by Hurst et al. {Hurst:2006ea} The AuNPs were

further stabilized using a thiol functionalized poly(ethylene glycol) methyl ether ligand,

for the same reason QDs were functionalized with mPEG ligands (see above). When

required, monovalent conjugates of both QDs and AuNPs were prepared using the

protocol developed by Uddayasankar et al. {Uddayasankar:2014fi} (See supporting

information, Figure S2).

In the subsequent discussion, the QDs are classified using their emission

wavelength. For example, QDs that have an emission wavelength of 525 nm are referred

to as QD525. The AuNPs are classified using their diameter. For example, an AuNP with

a diameter of 6 nm would be referred to as 6AuNP.

Configuration 1: One AuNP – multiple QD system

Forming QD-AuNP complexes in solution results in a decrease of the

fluorescence intensity from QDs. This occurs for two reasons: the inner filter effect and

non-radiative energy transfer.

Assessing the influence of an inner filter effect requires QDs and AuNPs that co-

exist in solution, but with no physical contact between the two nanoparticles. Both the

QD525s and the AuNPs of three different diameters (6, 13 and 30 nm) were coated with

mPEG based ligands to minimize any non-specific interactions. The concentration of

AuNPs was varied while keeping the QD concentration constant at 50 nM. This provided

for solutions with varying ratios of QDs to AuNPs and since QD concentration was held

constant, an increase ratio is indicative of decreasing AuNP concentration. An idealized

physical model based on a packing arrangement of spheres around a central sphere was

taken into consideration when determining the appropriate ratios of QD525 to AuNP for

the different sizes of AuNPs (Supporting information, Table S1). The different sizes of

AuNPs were expected to accommodate varying numbers of QDs, necessitating the

investigation of multiple ratios of QDs per AuNP.

The presence of any inner filter effect was assessed by measuring the fluorescence

intensity of the series of solutions relative to a solution of QDs with no AuNPs. The data

from these experiments are presented in Figure 2 for three different sizes of AuNP (a – 6

nm, c – 13 nm and e – 30 nm). For all sizes of AuNPs, the general trend indicates a

decrease in fluorescence intensity for higher concentrations of gold nanoparticle (lower

ratios of QD to AuNP). Depending on the spectral characteristics of the chromophore, the

decrease in fluorescence intensity due to an inner filter effect is a combination of both a

decrease in penetration depth of excitation intensity, and a decrease in the emitted

intensity from the fluorophores. The absorbance spectra of the AuNPs and the emission

spectrum of QD525 are available in Figure S4. AuNPs have a significant extinction

coefficient at 405 nm (wavelength used to excite QDs), resulting in the attenuation of the

excitation source intensity. In fact, even the smallest AuNPs used in these experiments

(6AuNP) have an extinction coefficient at 405 nm (1.3x107 M-1cm-1) that is almost 20

times greater than the extinction coefficient of QD525 (6.8x105 M-1cm-1) at that

wavelength. In addition, the high spectral overlap between the emission of QD525 and

the plasmon absorption peaks of the AuNPs further attenuate the light intensity emitted

by the QDs (for spectral overlap integrals, see Table S2 in the Supporting information).

The 30AuNPs have the most significant impact with regards to decreasing the observed

photoluminescence of QDs due to the inner filter effect. Even for the highest ratios of QD

to AuNP (i.e. lowest AuNP concentration) that were investigated, only 30 % of the

fluorescence intensity was observable. For the 6AuNP and 13AuNP systems, up to 80 %

of the fluorescence intensity was retained at the highest ratios of QD to AuNP. Since the

attenuated fluorescence intensity is due to a reduction in excitation radiation and emitted

radiation, the product of the transmittance at the excitation wavelength and emission

wavelength (of a solution of AuNPs) should provide an estimate of the inner filter effect.

Figure S5 in the Supporting Information serves to compare the fluorescence intensities

observed to those predicted by the transmittance of AuNP solutions. While an exact

quantitative agreement between the theoretical predictions and experimental observations

was not expected, the trends that were experimentally observed closely followed the

theoretical predictions. Although the mPEG coatings are expected to reduce the

interactions between the nanoparticles, {Choi:2012ja} deviations between theory and

experimental data could result from some non-specific interactions between the

nanoparticles. Furthermore, colloidal AuNPs often have a distribution in the size of

nanoparticles. A distribution in the size of AuNPs results in a distribution of extinction

coefficients of the nanoparticles, adding more uncertainty to the prediction of the inner

filter effect. The most accurate method to assess the influence of an inner filter effect is

by experimental quantification.{Zhang:2015fu}

Figure 2 Effect of gold nanoparticle size on the performance of displacement assays involving quantum dots as donors and gold nanoparticles as acceptors. The trends in the inner filter effect (left axis) and contrast ratio (right axis) as a function of number of QDs per AuNP for three different sizes (diameter) of AuNPs; (a) 6 nm (c) 13 nm and (e) 30 nm. The TEM images provide confirmation of QD binding onto AuNPs of the three different sizes: (b) 6 nm (d) 13 nm and (f) 30 nm.

By establishing the trends associated with the inner filter effect, it was then

possible to evaluate the influence of AuNP size on the contrast ratio. Contrast ratio was

used as a parameter to evaluate energy transfer efficiencies, and is defined as the ratio of

the fluorescence intensity after target binding to that prior to introduction of the

oligonucleotide targets. Incubation of the AuNP-DNA conjugates with monovalently

functionalized QD-DNA conjugates resulted in the formation of “satellite structures” as

confirmed with TEM imaging (Figure 2(b), (d) and (f)). The monovalency of the QDs

prevented the crosslinking of multiple AuNPs. Introduction of the target oligonucleotides

resulted in the disassembly of the AuNP-QD satellite structures. To minimize the

influence of variable hybridization efficiencies, a high equivalence of target (>100 equiv.

with respect to the QD-DNA monoconjugates) was used. The plots represented in Figure

2 demonstrate the trends in contrast ratio for different ratios of QDs per AuNP, with each

of the panels corresponding to a different AuNP diameter. The highest contrast ratios

were observed for the 13 nm AuNP system, and even the lowest contrast ratio observed

with this system was higher than the largest ratios observed for the other two sizes of

AuNPs.

To investigate the quenching efficiencies of the AuNP – QD conjugates, the

fluorescence lifetime of QDs hybridized to AuNPs of the different diameters were

compared to QDs in the absence of AuNPs. The lowest QD to AuNP ratio presented in

Figure 2 was used for all sizes of AuNPs to ensure minimal contribution from QDs that

were not bound to AuNPs. The fluorescence lifetime, and associated quenching

efficiencies are presented in Table 1.

Table 1 Fluorescence lifetimes and quenching efficiency of QD525 with three different gold nanoparticle sizes.

Sample Fluorescence lifetime (ns)

Quenching efficiencya (%)

QD525 (No AuNP) 5.3 ± 0.3 -

QD525 – 6AuNP 3.2 ± 0.2 40 ± 1

QD525 – 13AuNP 1.7 ± 0.2 68 ± 1

QD525 – 30AuNP 3.6 ± 0.4 32 ± 1

aQuenching efficiency=(1−( τ AuNP

τ0 ))×100; where τAuNP refers to the fluorescence lifetime in the

presence of AuNPs and τ0 refers to that in the absence of AuNPs

Compared to QDs alone, the QDs incubated with AuNPs have a statistically

significant decrease in fluorescence lifetime, as is expected in scenarios involving non-

radiative energy transfer.{Lakowicz:tx} Comparing the fluorescence lifetimes in the

presence and absence of quencher allows the calculation of energy transfer efficiency (or

quenching efficiency). The 13AuNPs had a higher energy transfer efficiency compared to

the 6AuNP. This was expected given that the spectral overlap integral of the 13AuNP is

approximately an order of magnitude greater than that of the 6AuNP (Supporting

information, Table S2). These results are consistent with previous studies where larger

AuNPs provide for a higher quenching efficiency {Griffin:2009wr}, and is also

supported by the higher contrast ratio observed for the 13AuNP – QD system.

The 30AuNP – QD system had the lowest energy transfer efficiency of 32 %,

even though it has a spectral overlap integral that is 24 times greater than that of the

13AuNP (Table S2). This lower energy transfer efficiency is also consistent with the

lower contrast ratio observed for the 30AuNP – QD system. This issue was further

investigated by analyzing the hybridization efficiency between the AuNPs and QDs. This

was done only for the 13AuNP - QD and 30AuNP – QD system to resolve the

inconsistencies seen in the quenching efficiency measurements. Taking into account the

theoretical maximum of the number of QDs that may be assembled onto the AuNPs (See

Supporting Information, Table S1), the 13AuNP and 30AuNP were first incubated with

varying equivalents of QD525-DNA monoconjugates. The samples containing the two

nanoparticles were centrifuged at a sufficient centrifugal force to precipitate the AuNPs

while keeping the QDs in solution. Comparison of the quantity of QDs remaining in

solution to the initial amount of QDs provided an estimate of the number of QDs bound

to the AuNPs (Figure 3).

Figure 3 Hybridization efficiency of QD525 – probe DNA monoconjugates onto gold nanoparticles functionalized with displacer DNA. (a) 13 nm AuNP (b) 30 nm AuNP.

As the number of QDs incubated with 13AuNP increased, a decrease in the

hybridization efficiency was observed (Figure 3a). A lower hybridization efficiency at

QD to 13AuNP ratios greater than 10 is consistent with the trend of decreasing contrast

ratio as the ratio of QD to AuNP increases (Figure 2c). The QDs not bound onto the

13AuNP contributed to the fluorescence intensity measured. This provides higher

background fluorescence regardless of the presence of target, leading to a smaller

contrast ratio. The lack of quantitative binding, even for the smallest ratios of QD525 to

13AuNP, can be attributed to the equilibrium that exists between the bound and free form

of the QD-DNA monoconjugates. Decreasing the concentration of 13AuNP (i.e.

increasing ratio of QD to AuNP) results in a shift in the equilibrium to the unbound state,

consistent with the lower hybridization efficiencies observed at higher ratios of QD to

13AuNP. The fixed geometric volume around the AuNPs also limits the total number of

QDs that may be immobilized per AuNP.

Figure 3b illustrates the hybridization efficiencies of monovalent QD-DNA

conjugates on 30AuNPs. A decrease in the hybridization efficiency was noted as the ratio

of QDs to 30AuNP was increased as also observed for the 13AuNP-QD assemblies, but

the reduction of hybridization efficiency was more pronounced for the 30AuNPs. As the

size of an AuNP increases, the curvature of the surface decreases. A direct consequence

of surface curvature is the divergence angle between immobilized oligonucleotides.

Using the divergence angles estimated by Hill et al {Hill:2009vr}, 25-mer

oligonucleotides on 30AuNPs should have a divergence angle of 140, while those on the

13AuNPs have a divergence angle of approximately 240. The reduced divergence angle

on the 30AuNPs could lead to greater steric interactions between neighboring

biomolecules, shifting the equilibrium to the unbound state of QD-DNA monoconjugates.

Steric limitations are particularly important considerations in this system due to proximal

positioning of the QD and AuNP, and could explain the low hybridization efficiencies on

the 30AuNPs. The lower hybridization efficiency of the 30AuNPs results in a higher

background signal from QDs that remain in bulk solution (prior to target introduction).

This has the consequence of lowering the contrast ratio, even though larger nanoparticles

are known to have a higher quenching efficiency. These observations reflect the trends

observed in the contrast ratio as a function of AuNP size (Figure 2), and are consistent

with the fluorescence lifetime data (Table 1), for which the smaller decrease in lifetime

was attributed to the presence of QDs not hybridized onto the AuNPs.

It should be noted that purification of AuNP-QD complexes from unhybridized

QD-DNA conjugates was attempted by centrifugation. However, centrifugation

negatively impacted the colloidal stability of the AuNP-QD complexes, and significant

sedimentation was observed. The response of the purified complexes to the presence of

target DNA strands was very low (results not shown).

The DNA sequences were designed to ensure that the QDs and AuNPs were on

the same end of the DNA hybrid. Investigation of the impact of proximal versus distal

positioning of QDs was completed to evaluate any impact of steric hindrance on surface

loading (Figure 4).

Figure 4 (a) Schematic illustration of the proximal and distal configuration of the nanoparticle complex for nucleic acid displacement assays (b) Contrast ratio of distal and proximal configurations for AuNPs of three different sizes.

The contrast ratios obtained for the three different sizes of AuNPs for both the

proximal and distal configurations are listed in Figure 4b. For all sizes of AuNPs, the

contrast ratio was higher for the proximal than the distal configuration. The proximal

design positions the QD closer to the AuNP surface, improving non-radiative energy

transfer efficiency. Figure S6 in the supporting information illustrates the absolute

fluorescence intensities for the QD-AuNP conjugates before and after the introduction of

target. For the distal and proximal configurations, the differences observed in the

fluorescence intensities after the target DNA strand displaces the QDs from the AuNPs is

minimal (<5 %), while the differences in fluorescence intensity prior to target

introduction are greater. The difference is greatest for the 13AuNP system, where the

distal configuration has a 28 % greater fluorescence intensity when compared to the

proximal configuration. These observations suggest that the variations in contrast ratios

between proximal and distal configurations arise predominantly from the lower

quenching efficiency in the distal configuration. The differences are not as significant for

the 6AuNP and 30AuNP system due to their initial low contrast ratios. These

observations demonstrate the lack of any detrimental effects to placing the nanoparticles

in a proximal arrangement.

Contrast ratio and inner filter effects were also determined for QDs that emit at

575 nm and 630 nm (Figure 5). Monovalent conjugates of QDs with emission

wavelengths of 575 nm and 630 nm were prepared by the same protocol used for the

QD525. Complexes of QDs and AuNPs were prepared using only 13AuNP so they may

be evaluated with the optimal performance previously noted for QD525. Emission spectra

overlaid onto the absorption spectra of 13AuNPs are provided in the supporting

information (Figure S7)

Figure 5 Effect of QD emission wavelength on displacement assay performance where QDs serve as the donor and 13AuNPs serve as the acceptor, with multiple QDs arranged around the AuNP (Configuration 1). The trends in inner filter effect (left axis) and contrast ratio (right axis) as a function of number of QDs per AuNP for AuNPs of two different emission wavelengths (a) 575 nm (b) 630 nm.

A greater fraction of fluorescence intensity is preserved for the QD575 and

QD630 compared to the QD525 (Figure 2c) when considering the inner filter effect. This

may be rationalized by the smaller spectral overlap integral between the emission

intensities of the QDs and the 13AuNP (Table S3). The higher fluorescence intensities

were accompanied by a lower contrast ratio, and the results are consistent with the

decrease in the spectral overlap between the emission of the QDs and the absorbance of

13AuNPs.

Configuration 2: One QD – multiple AuNP system

The configuration most commonly used in assay development makes use of a

single QD surrounded by multiple AuNPs that serve as acceptors. Such a configuration

takes advantage of higher energy transfer efficiencies obtained by surrounding a single

donor (QD) with multiple acceptors (AuNPs). QD525 was functionalized with multiple

DNA strands, with successful conjugation confirmed using agarose gel electrophoretic

analysis and UV-vis spectroscopy (Figure S8). 6AuNPs were monofunctionalized with

DNA using the same protocol developed for QDs (see Supporting Information, Figure

S2). Assembly of the conjugates was accomplished by incubating the DNA modified

QDs and AuNPs at specific ratios. 6AuNP, rather than smaller AuNPs, was chosen as the

acceptor to provide for improved analytical performance.{Choi:2012ja} In these

experiments the AuNP concentration was kept constant at 50 nM, with the amount of

QDs varied to obtain the specified AuNP to QD ratios. It is important to note that in

Configuration 1, it was the QDs that were kept at a constant concentration, with the

concentration of AuNP varied to obtain the desired ratios. In Configuration 1, it was the

QDs that hybridized to the target. In Configuration 2, the target hybridized to the 6AuNP.

To make a valid comparison between the two configurations, the concentration of

nanoparticles that responds to the target DNA should be the same. Contrast ratios were

obtained by taking fluorescence measurements before and after incubation of the

complexes with an excess of target DNA (>100 equiv.). Figure 6 depicts the contrast

ratios for varying ratios of AuNP to QD.

Figure 6 Inner filter effect (left axis) and contrast ratio (right axis) as dependent on the number of AuNPs per QD, for a displacement assay system involving multiple AuNPs interacting with a single QD (Configuration 2).

Increasing the ratio of AuNPs to QD resulted in an increase in the contrast ratio.

For assays based on energy transfer, a greater number of acceptors per donor improves

energy transfer efficiency. {Medintz:2003fr} This provides for a greater quenching

efficiency of the QDs, leading to a larger contrast ratio. For equivalences <2, a contrast

ratio of less than 1.5 is obtained. Equivalences <2 should be avoided due to the presence

of QDs that are not associated with an acceptor AuNP. The high surface area of the QDs

results in a statistical distribution of QD-AuNP conjugates that can impact assay

performance, especially at low equivalences. {Scott:2014iv} Assuming the distribution

follows Poisson statistics, an AuNP to QD equivalence of 1 leaves 37 % of the QDs

without an acceptor, and this fraction decreases to 14 % for an equivalence of 2. QDs that

have no acceptor would contribute to the background signal without taking part in the

assay. Above an equivalence of two, contrast ratios greater than 2 are obtained with 5

AuNPs per QD providing a contrast ratio of approximately 3. Equivalences above 5 were

not attempted because the spectrofluorimeter could not detect the fluorescence intensity

of the QDs in solutions for AuNP to QD ratios greater than 5.

The inner filter effect was also investigated (Figure 6). Both QDs and AuNPs

were functionalized with mPEG ligands to reduce non-specific interactions. Increasing

the ratio of AuNP to QD leads to a decrease in the observed fluorescence intensity, yet

the correlation between the decrease in fluorescence and the ratio of AuNPs per QD is

not as pronounced as that observed for Configuration 1 (Figure 2). This is attributed to

the fact that the concentration of AuNPs was kept constant for the different solutions. The

constant concentration of AuNPs attenuated the excitation and emitted radiation to the

same extent in solutions with varying ratios of AuNP to QD. For the ratios of AuNP to

QD investigated, only 15 - 26 % of the fluorescence was observed (when compared to a

situation with no AuNPs).

Comparing Configuration 1 and Configuration 2

The analytical performance of the two configurations was investigated by looking

at the response of the systems to increasing target concentrations (Figure 7).

Figure 7 Comparing the analytical performance of the two configurations of displacement assay. (a) The contrast ratio as a function of target concentration for the two different assay configurations, each performed at two different ratios of the QDs and AuNPs. (b) Absolute fluorescence intensities of the sample solutions corresponding to the different assay configurations.

Figure 7a provides the response curves for four different nanoparticle

configurations (two representative examples of Configuration 1, two of Configuration 2).

All arrangements had the same number of monovalent nanoparticles, and this is reflected

in the response curves where the response saturated using the same amount of target.

Presented in Figure 7b are the absolute fluorescence intensities of the solutions for the

four different arrangements. The variations in the observed fluorescence intensity arise

from the differential inner filter effects of the various nanoparticle arrangements. To

determine the optimal arrangement, both fluorescence intensity and any inner filter effect

need to be considered concurrently.

The response curves with the lowest and highest responses to changing quantity

of target correspond to Configuration 2 with two 6AuNP per QD525 and Configuration 1

with 5 QD525 per 13AuNP, respectively. This particular arrangement of Configuration 1

had a greater fluorescence intensity than that of the Configuration 2, which is a

consequence of the lower inner filter effect observed when multiple QDs are associated

with a single AuNP, rather than multiple AuNPs being associated with a single QD.

The other two response curves presented in Figure 7a correspond to Configuration

1 with 15 QD525 per 13AuNP and Configuration 2 with three 6AuNP per QD525. Both

these curves show similar sensitivities to changing amounts of target, but the observed

fluorescence intensities are significantly different between these configurations. The

intensities of the Configuration 1 arrangement were approximately five times greater than

for the construct based on multiple 6AuNPs arranged around a single QD.

An optimal configuration should provide high absolute fluorescence intensities

(i.e. low inner filter effect) while at the same time offering a high contrast ratio to allow

for sensitive measurements. This is certainly desirable when assays require the use of

simplified readout platforms such as consumer cellphone/tablet cameras. Using these

criteria, the configuration based on 15 QD525s around a single 13AuNP is optimal

(Figure 8).

Figure 8 Fluorescence images collected using an iPad camera for QD-AuNP complexes incubated with varying amounts of target (pmol) (i) 0 (ii) 0.20, (iii) 0.30, (iv) 0.40, (v) 0.50, (vi) 0.60, (vii) 0.70, (viii) 0.80, (ix) 1.00, (x) 2. (a) 2 6AuNP per QD525 (b) 3 6AuNP per QD525 (c) 15 QD525 per 13AuNP (d) 5 QD525 per 13AuNP (e) Response curves obtained from images in panel (a) and (b). (f) Response curves obtained from images in panel (c) and (d).

Fluorescence photographs (using UV lamp excitation) of the solutions used to

obtain the response curves depicted in Figure 7 are shown in Figure 8. Visual inspection

of the images shows the significant improvement in assay performance that may be

obtained from Configuration 1 (Figure 8c and d), as opposed to Configuration 2 (Figure

8a and b). Fluorescence is difficult to visually discern in panels (a) and (b), while the

fluorescence intensities are obvious in panels (c) and (d). Intensity measurements

determined using a spectrofluorimeter (Figure 7) demonstrated that Configuration 2

provided similar sensitivities when compared to Configuration 1 (Table 2). However, the

use of a tablet camera to collect a fluorescence image using Configuration 2 provided

data that was of low quality for construction of a response curve (Figure 8e). The use of

Configuration 1 provided images that yielded a response curve that were better suited for

quantitative analysis (Figure 8f).

Table 2 Response curve sensitivity for Configuration 1 and Configuration 2

Configuration Slope ( x 10-13 )aSpectrofluorimeter iPad Camera

5 QD525 per 13AuNP 2.8 ± 0.7 2.1 ± 0.815 QD525 per 13AuNP 2.7 ± 0.4 2.3 ± 0.63 6AuNP per QD525 2.1 ± 0.5 0.1 ± 1632 6AuNP per QD525 1.8 ± 0.4 0.2 ± 0.7

aThe slope was determined from the equation of the sigmoidal curve for the different

dose-response plots: Contrast ratio=base+ max

1+e( xhalf −x

slope )

This is further supported by sensitivity measurements of the various response

curves, which is presented in Table 2. Using the spectrofluorimeter to measure the

samples provided similar sensitivities for all configurations, with Configuration 1 having

slightly higher sensitivities (at most 25 % better). But looking at the data collected using

the iPad camera, Configuration 2 had sensitivities that were an order of magnitude lower

than those for Configuration 1. In addition, the high uncertainties associated with the

slope of the response curves obtained by the iPad camera is indicative of poor fit of the

data to the expected sigmoidal curve shape. Comparing the two configurations,

Configuration 1 provides higher signal intensities while still maintaining the assay

response to target hybridization (contrast ratio).

Conclusion

The relative arrangement of QDs and AuNPs has a significant impact on the

analytical performance of assays that rely on energy transfer between the two

nanoparticles. Using monovalently functionalized nanoparticles, QD-AuNP complexes

were prepared in two different configurations. Configuration 1 assembled monovalent

QD-DNA conjugates around a central AuNP and made use of three different sizes of

AuNPs (6 nm, 13 nm and 30 nm). Configuration 2 placed monovalent 6AuNP-DNA

conjugates around a central QD. Both these configurations were assembled using DNA

hybridization. Energy transfer processes were investigated by looking at fluorescence

intensity changes as the complexes were disassembled using DNA strand displacement

reactions. Additionally, the strong optical absorption characteristics of AuNPs led to

significant reductions of fluorescence intensity due to the inner filter effect. Both these

factors were concurrently evaluated to establish the optimal configuration. Optimum

performance was obtained using Configuration 1 with at most 15 QDs assembled around

a 13AuNP. This system offered a 2.5 fold increase in fluorescence signal upon target

recognition. Similar increases in signal intensity using Configuration 2 required at least

three 6AuNP around the QD. However, this resulted in Configuration 2 having a 5 fold

lower signal intensity (compared to Configuration 1) due to inner filter effects. Dose-

response curves were also constructed for the two configurations, the sensitivities of

which indicated that Configuration 1 provided better analytical performance. In fact,

when low cost detectors were used (digital camera), Configuration 2 had sensitivities that

were an order of magnitude lower than those obtained using Configuration 1. These

results serve to provide important guidelines for the design of assays that use gold

nanoparticles and quantum dots as part of an energy transfer assay.