Article pubs.acs.org/acssensors
Evaluating Dye-Labeled DNA Dendrimers for Potential Applications in Molecular Biosensing Carl W. Brown, III,†,∥ Susan Buckhout-White,† Sebastián A. Díaz,†,§ Joseph S. Melinger,‡ Mario G. Ancona,‡ Ellen R. Goldman,† and Igor L. Medintz*,† †
Center for Bio/Molecular Science and Engineering, Code 6900, and ‡Electronics Science and Technology Division, Code 6876, U.S. Naval Research Laboratory, Washington, DC 20375, United States § American Society for Engineering Education, Washington, DC 20036, United States ∥ College of Science, George Mason University, Fairfax, Virginia 22030, United States S Supporting Information *
ABSTRACT: DNA nanostructures provide a reliable and predictable scaffold for precisely positioning fluorescent dyes to form energy transfer cascades. Furthermore, these structures and their attendant dye networks can be dynamically manipulated by biochemical inputs, with the changes reflected in the spectral response. However, the complexity of DNA structures that have undergone such types of manipulation for direct biosensing applications is quite limited. Here, we investigate four different modification strategies to effect such dynamic manipulations using a DNA dendrimer scaffold as a testbed, and with applications to biosensing in mind. The dendrimer has a 2:1 branching ratio that organizes the dyes into a FRET-based antenna in which excitonic energy generated on multiple initial Cy3 dyes displayed at the periphery is then transferred inward through Cy3.5 and/ or Cy5 relay dyes to a Cy5.5 final acceptor at the focus. Advantages of this design included good transfer efficiency, large spectral separation between the initial donor and final acceptor emissions for signal transduction, and an inherent tolerance to defects. Of the approaches to structural rearrangement, the first two mechanisms we consider employed either toehold-mediated strand displacement or strand replacement and their impact was mainly via direct transfer efficiency, while the other two were more global in their effect using either a belting mechanism or an 8-arm star nanostructure to compress the nanostructure and thereby modulate its spectral response through an enhancement in parallelism. The performance of these mechanisms, their ability to reset, and how they might be utilized in biosensing applications are discussed. KEYWORDS: dendrimer, DNA, FRET, nanostructure, dye, energy transfer, biosensing
O
energy transfer and light-harvesting, with the motivation being a perceived potential for exploitation in biosensing, energy conversion, biocomputing, cryptography, data storage, plasmonics, and so forth.6−14 In this type of work, conjugated biomolecules (i.e., organic fluorophores) and other optically active materials (e.g., quantum dots, fluorescent proteins, rare earth metal cryptates, bioluminescent enzymes) are typically arranged in precise conformations within and around the DNA structure to give rise to Förster resonance energy transfer (FRET) cascades.13,15,16 In addition to light-harvesting and
ver the past decades, the use of DNA as a biocompatible material for the engineering of static structures and dynamic scaffolds within synthetic biology and bionanotechnology has been widely explored.1−3 DNA has a number of significant advantages that make it an excellent structural material at the nanoscale: Its foundation in Watson−Crick base pairing gives rise to a straightforward design process with readily predictable results; it is easily modified with other chemical species such as fluorescent dyes; its natural role as an information carrier provides a direct path for this information to be stored, read, interpreted, and transmitted, often in the form of logic gates; and its dynamic capabilities are well suited for biosensor development.4,5 One area where the application of DNA nanostructures is growing quite rapidly is that of excitonic © XXXX American Chemical Society
Received: December 1, 2016 Accepted: February 23, 2017
A
DOI: 10.1021/acssensors.6b00778 ACS Sens. XXXX, XXX, XXX−XXX
Article
ACS Sensors
Figure 1. Schematic detailing of the various structural modification approaches to alter FRET in the DNA dendrimer. The parent 2:1 donor:acceptor DNA dendrimer nanostructure (0) acts as a scaffold for the precise positioning of a four-dye three FRET step cascade. Dyes are represented by colored stars, Cy3 (green, 8), Cy3.5 (yellow, 4), Cy5 (orange, 2), and Cy5.5 (red, 1). Four approaches were considered, two targeting internal structural modifications and two that externally modify the DNA dendrimer: 1, Strand Displacement - an invader strand binds to and removes the Cy5 labeled strand (orange) from the core dendrimer structure; 2, Strand Replacement - an invader strand binds to the Cy3 toehold (green) and competes off the Cy5 strand (orange), replacing it within in the dendrimer; 3, Belted dendrimer - two belt strands bind to the toehold (bypassing Cy3 toeholds, green), constricting the structure by closing its circumference; 4, 8-arm Star - a separate nanostructure that binds alternating toeholds, constricting the base dendrimer into either a flat or a ball-like form. Added DNA strands or DNA structures are depicted in gray. The predicted changes to the FRET pathways, where arrows indicate FRET steps, include either altering FRET (red X) or increasing overall FRET efficiency through the cascade (red arrow).
found to be the presence of multiple overlapping energy transfer pathways for excitons to traverse the structure; this also helped compensate for any structural deficiencies such as missing or photobleached dyes. Although several different dendrimeric structures were investigated, the 2:1 branching architecture showed the most versatility, manifesting efficient FRET between each step while minimizing structural complexity and formation issues (Figure 1). While these dendrimers gave important insight into the design and operation of DNA-based cascaded-FRET nanostructures, the earlier work focused only on the static regime. Here we extend from this to consider their dynamic behavior in which altering or rearranging the underlying structure is transduced into a modified spectral response. The goal here is to evaluate their performance in this modality and especially their potential benefits and liabilities for biosensing applications.
directed energy transfer, dye-labeled DNA structures also form the backbone of DNA-based sensors, with the most common iteration being the molecular beacon (MB).17−19 Although widely implemented, many beacon designs are insufficient for multiplexed sensing formats, where multiple donor−acceptor pairs are present and target different sequences. This arises due to spectral cross-talk between closely emitting dyes in the visible portion of the spectrum along with direct excitation of the acceptor(s) multiple overlapping absorption and emission spectra.14,20 Moreover, linear single oligonucleotide MBs are not suited to sensing the presence of other complex DNA structures. As DNA structures grow in structural and functional complexity, the need for sensing different types of structures along with the ability to undertake multiple parallel or consecutive sensing events in complex formats will only grow. Potential ways to compensate for these two issues include adding an antenna to the MB to increase its “effective” Stokes shift and increase its dynamic signal range along with creating DNA sensors that are capable of engaging in complex interactions with other multimeric DNA structures. We recently examined the photophysical properties of a series of programmable photonic networks based on triangular, bifurcated, Holliday junction, 8-arm star, and dendrimeric DNA designs involving up to five different dyes engaged in four consecutive FRET steps.21,22 The dendrimeric structures were found to be particularly advantageous for assembling FRET components with exceptional enhancement of the end-to-end energy transfer efficiency (E−Eeff) and antenna gain. The mechanism underlying efficient FRET in the dendrimers was
■
EXPERIMENTAL SECTION
DNA and Structural Assembly. A detailed description is provided in the Supporting Information (SI). Photophysical and FRET Analysis. Spectral data was collected on a Tecan Infinite M1000 dual monochromator multifunction plate reader equipped with a xenon flash lamp (Tecan, Research Triangle Park, NC) using 515 nm excitation with data recorded over an emission range of 530−800 nm. Samples included all structures along with controls containing partial complements along with each individual dye to estimate the direct excitation contributions. Spectra were decomposed into their individual components and the direct excitation subtracted to yield curves reflecting the net donor fluorescence loss or acceptor sensitization.21−25 Spectral overlap functions or J(λ) and Förster B
DOI: 10.1021/acssensors.6b00778 ACS Sens. XXXX, XXX, XXX−XXX
Article
ACS Sensors
Figure 2. Averaged spectra following either targeted strand displacement or replacement. (A) Strand displacement of the Cy3.5-labeled DNA strand. (B) Strand displacement of the Cy5-labeled DNA strand. (C) Strand replacement of the Cy3.5-labeled DNA strand with an unlabeled DNA strand. (D) Strand replacement of the Cy5-labeled DNA strand with an unlabeled DNA strand. In all cases, a range of molar ratios (1×, 2×, 5×, and 10×, with 0× being the control) were used for the reactions as indicated. The inset represents the ratio of the PL area of the Cy5.5 curve over the area of the Cy3 PL curve. The color of the point corresponds to the color of the curve and ratio in the decomposed spectra. A trend line is added as a guide to the eye.
■
distances (R0) were estimated as before using standard FRET formalism.9,20,21,24 The average energy transfer efficiency E was extracted for each donor−acceptor pair using the expression:
E=
(FD − FDA ) FD
RESULTS Dendrimer Structure and Function. The core structure, whose modification is explored here, was described in BuckhoutWhite et al.24 and consists of a four-dye FRET cascade configured as a dendrimer placing a single Cy5.5 dye at its center with a 2:1 branching ratio extending outward from there and displaying a dye sequence of 2-Cy5, 4-Cy3.5, and 8-Cy3 dyes at the periphery (Figure 1). This arrangement is a functional compromise and represents a light-harvesting nanoantenna that achieves improved energy transfer efficiency while not being overly complex. In the original report, we found it to exhibit an E−Eeff of 28% and an antenna gain of 3.5 assuming a ∼70% formation efficiency.24 E−Eeff dropped to 17% when a single-point attached Cy3.5 ester dye was utilized within the structure instead of a more constricted Cy3.5 double-phosphate linkage.27,28 The dendrimer designs examined herein used the same dyes and had similar formation yields (60−70%, data not shown) and transfer efficiencies (vide infra). The structure was designed to position the dyes with an approximate spacing between each nearestneighbor donor−acceptor pair of ∼0.5× the Förster distance (R0, Table S5). According to Förster theory, this should yield singlestep FRET efficiencies (E, from a given donor to a given acceptor) of ≥90%, although such numbers are almost never realized in practice in unpurified samples.24 SI Figure S1A displays the excitation spectra of the dyes used, scaled by the extinction coefficient; Figure S1B displays emission spectra scaled by the QY. Figure S1C plots the spectral overlap functions between nearest-neighbor dyes confirming the basis for efficient FRET. Since all the dyes in these structures are placed relatively
(1)
where FD and FDA are, respectively, the fluorescence intensities of the donor alone and the donor in the presence of the acceptor(s).9,20 We define exciton transfer efficiency as the conditional probability that an exciton will transfer from a given excited donor to a given acceptor, with the process of most interest being the end-to-end transfer efficiency (E− Eeff) from a peripheral donor to a terminal acceptor which is estimated using: E − Eeff =
(ΦA − Φ0A )/Q A Φ0D /Q D
(2)
where ΦA−ΦA0 is the decomposed area under the emission or photoluminescence (PL) curve of the ending acceptor minus the direct excitation of that acceptor and Φ0D the PL area of the initial donor, Cy3, without any acceptors present. QA and QD are the quantum yields (QYs) of the acceptor and donor, respectively.23,25,26 In this expression, the denominator gives the number of excited donors while the numerator is the number of subsequently sensitized acceptors that did not become excited as a result of direct excitation. In cases where the acceptor to donor ratio is used in lieu of E−Eeff, that is, simply the area of the decomposed Cy5.5 acceptor, divided by the area of the Cy3 donor, again decomposed from the raw spectrum and corrected for direct excitation. C
DOI: 10.1021/acssensors.6b00778 ACS Sens. XXXX, XXX, XXX−XXX
Article
ACS Sensors
(∼10-fold) and a concomitant decrease in terminal Cy5.5 sensitization (∼5-fold) in proportion to the added complementary strand (Figure 2A). The inset in Figure 2A plots the ratio of the deconstructed peak areas of Cy5.5/Cy3 and shows a drop in value from around 1.6 to
