Triangulating Nucleic Acid Conformations Using Multicolor Surface Energy Transfer Ryan A. Riskowski,† Rachel E. Armstrong,‡ Nancy L. Greenbaum,§ and Geoffrey F. Strouse*,†,‡ †
Molecular Biophysics Program, Florida State University, Tallahassee, Florida 32306, United States Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Florida 32306, United States § Department of Chemistry and Biochemistry, Hunter College and The Graduate Center of the City University of New York, New York, New York 10065, United States ‡
S Supporting Information *
ABSTRACT: Optical ruler methods employing multiple fluorescent labels offer great potential for correlating distances among several sites, but are generally limited to interlabel distances under 10 nm and suffer from complications due to spectral overlap. Here we demonstrate a multicolor surface energy transfer (McSET) technique able to triangulate multiple points on a biopolymer, allowing for analysis of global structure in complex biomolecules. McSET couples the competitive energy transfer pathways of Förster Resonance Energy Transfer (FRET) with gold-nanoparticle mediated Surface Energy Transfer (SET) in order to correlate systematically labeled points on the structure at distances greater than 10 nm and with reduced spectral overlap. To demonstrate the McSET method, the structures of a linear B-DNA and a more complex folded RNA ribozyme were analyzed within the McSET mathematical framework. The improved multicolor optical ruler method takes advantage of the broad spectral range and distances achievable when using a gold nanoparticle as the lowest energy acceptor. The ability to report distance information simultaneously across multiple length scales, short-range (10−50 Å), mid-range (50−150 Å), and long-range (150−350 Å), distinguishes this approach from other multicolor energy transfer methods. KEYWORDS: surface resonance energy transfer, Förster resonance energy transfer, noble metal nanoparticles, nucleic acid structure modeling, long-range optical molecular ruler
M
Here we demonstrate the ability to measure three structural distances within a biomolecule simultaneously following excitation of a bound molecular dye by multicolor surface energy transfer (McSET). The McSET technique couples FRET, which operates in the short- (10−50 Å) to mid-range (50−150 Å) distance scale, and SET, which operates in the mid- to long-range (50−350 Å) scale. As a result, this technique permits multiple length scales to be monitored simultaneously in a single biomolecule. The energy coupling mechanism in small particle SET leads to an R−4 distance-dependent behavior, which is fully described in the literature as an enhancement of the radiative and nonradiative pathways of the molecular dye within the near field of the nanometal.14−18 In contrast, FRET operates on a much shorter length scale and follows an R−6 distance dependence that is described by a dipole−dipole exchange mechanism.19 Combining FRET and SET within the
apping the complex conformational changes that occur globally in biomolecules, whether for a DNA duplex or the more complex folding seen in RNA ribozymes or riboswitches, is integral to understanding structure−function relationships in nature.1−5 Conformational changes resulting in global structural change are associated with binding of proteins to nucleic acids or Mg2+ to a self-splicing intron, formation of g-quadruplexes at a replication site, or folding of the telomere at the end of a chromosome, to name a few select examples.6−13 Accurate measurement and correlation of conformational changes occurring across multiple distances (i.e., short-range (10−50 Å), mid-range (50−150 Å), and longrange (150−350 Å)) are not straightforward, particularly at typical biological concentrations. There is great demand to develop methods to probe conformational motion. Examples such as Förster Resonance Energy Transfer (FRET) or goldnanoparticle mediated Surface Energy Transfer (SET) use fluorescent labels on a biopolymer and provide conformational insight over the entire structure. © XXXX American Chemical Society
Received: September 13, 2015 Accepted: January 21, 2016
A
DOI: 10.1021/acsnano.5b05764 ACS Nano XXXX, XXX, XXX−XXX
Article
www.acsnano.org
Article
ACS Nano
(Scenario I) and a folded RNA Type I−III Hammerhead ribozyme (Scenario II).46 These two examples were chosen to explore the versatility of using competitive SET and FRET pathways in a single system for analysis of the global structure of a biopolymer. In Scenario I, the linear strand of duplex DNA is internally labeled with the fluorophore ROX, which acts simultaneously as a FRET donor to DyLt680 and as a SET donor to a 10 nm gold nanoparticle (AuNP). In this 3-label system, DyLt680 (FRET acceptor) and the 10 nm AuNP (SET acceptor) are coupled to the ROX donor to provide competitive energy transfer pathways for the decay of donor excitation. To minimize contributions from interactions with DyLt680, the AuNP is appended to the DNA at a distance sufficiently far from the DyLt680 label, so that only ROX is quenched by the AuNP. The Hammerhead ribozyme structure in Scenario II represents a more complex labeling strategy in which all three of the labels are simultaneously coupled, and a 2 nm AuNP acts as a SET acceptor for both dyes of the FRET pair (Cy3AF647). Judicious selection of two different gold particle sizes in Scenario I and Scenario II demonstrates the flexibility within McSET to probe individual distances by choice of dye pair and AuNP size. Use of a 10 nm AuNP in Scenario I shows the potential for measuring large interlabel distances (>10 nm), and has the benefit of revealing the reliability of McSET even when the scattering component is measurable for particles in this size range. In Scenario II, we are able to probe contributions to energy transfer rates from the nanoparticle’s plasmonic near field by taking advantage of the close interlabel distances associated with the Hammerhead structure. Bringing the dye labels into close proximity to the nanoparticle surface increases the relative magnitude of the near field impinging on the dye label, but can also severely quench the radiative output and reduce the measurable signal. Selecting a 2 nm particle with reduced quenching efficiency ensures that the multiplexed FRET and SET pathways can be observed clearly for the label− label distances presented. Optical triangulation of the contact distances between the dyes and the AuNP in the two Scenarios is accomplished by fitting the experimentally measured FRET-SET lifetime of the emissive dyes measured in the picosecond time domain, including rise times for the FRET donor−acceptor pairs in the presence of the AuNP. The measured distances from the two Scenarios are compared with theoretical distances extracted from relevant structure models (see Materials and Methods) for evaluating the quality of fit. It is important to the McSET study that results of the optical experiments demonstrate that energy coupling arises from the lowest energy excited state of the donor in both the SET and FRET pathways, thus allowing the resonant exchange coupling to the FRET acceptor and SET metal acceptor to be treated as a simple kinetic model reflecting the competitive energy coupling pathways. Because of the broad spectral overlap with the metal plasmon frequency, the McSET method is applicable to a wide spectral range of FRET pairs and broadens the biophysical structural toolset. The McSET technique should be implementable to ALEX schemes or single molecule detection platforms.
McSET model thus enables concurrent monitoring of multiple distance ranges following excitation at a single wavelength. Distances between labels are triangulated by McSET analysis of competing energy transfer pathways between dyes. While other groups have reported the use of combined SET and FRET optical probes to observe the near-field effects of large nanoparticles (>20 nm),20−29 the application in a ruler method to map complex landscapes has not appeared. The most common tools of choice for structural biophysics include NMR, X-ray crystallography, and cryogenic electron microscopy (cryo-EM), which often reach resolutions ≤3 Å. However, these approaches typically require large quantities of material and are not routine techniques. To propose a structural model based upon NMR or cryo-TEM data, complex modeling and ordered systems are required.30−34 Optical methods based on energy transfer are now fairly routine in biophysics, reflecting the readily available selection of dyes, and can measure separation distances between two positions on a biopolymer with high precision, even at the single molecule level.19,35,36 In addition to FRET, which measures distances between 50 and 100 Å, plasmonic rulers and single-color SET methods have also been reported in the past decade to expand the range of the biophysical toolset, but have been traditionally limited to reporting single point contacts and have not yet translated to structural study of biopolymers.10,36,43−45 To measure multiple contact distances within a biopolymer, the dye pairs must be systematically moved, requiring multiple experiments. Although multicolor FRET has used three- or four-color labels simultaneously,37−39 this technique is still limited to short contact distances, is complicated by undesired excess energy transfer pathways, and suffers from inherent spectral overlap of the dye pairs. A multicolor technique that holds promise is the use of nested FRET pairs on a single structure that are independently excited, as described by the ALEX technique. The ALEX technique reduces the number of competitive energy coupling pathways, but still does not allow single experiment detection across multiple length scales and multiple points, and does not avoid the pitfall of spectral cross-talk, which can limit the precision of the global structural analysis.37,40 An alternative two-color strategy uses a nonemissive organic dye, or “dark-quencher” to map energy transfer between two emitting dyes, thus effectively reducing spectral cross-talk; however, this approach still fails to provide significant improvement in measurable distances. Additionally, darkquenchers such as Black Hole Quencher can suffer from erratic blinking under normal experimental conditions.41,42 Alternatively, optical methods that use gold nanoparticles as dark quenchers have the advantage of significantly extending the limits of measurable interlabel distances while offering reduced spectral cross-talk when coupling to multiple dyes; additionally, they are resistant to photodegradation over the time course of an experiment. Plasmonic particles also have extinction coefficients that are orders of magnitude larger than those of molecular dark-quenchers; consequently, energy transfer to plasmonic particles takes place over a much greater distance. Coupling of SET- and FRET-sensitive dyes within a single biomolecule, by enabling short- and long-range distances to be simultaneously measured, provides a means to triangulate more complex structures by mapping multiple contact points within a biopolymer in a single excitation experiment. In this work, the utility of the McSET method is demonstrated by triangulating structures for two scenarios: a simple linear B-DNA duplex
RESULTS The two constructs designed to evaluate the ability to triangulate distances via McSET through competitive SET and FRET pathways are shown schematically in Figure 1. The B
DOI: 10.1021/acsnano.5b05764 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano
Figure 1. (A) Schematic images of the simultaneous energy coupling pathways among multiple optical labels positioned on a nucleic acid conformer in (Scenario I) a 60-mer duplex DNA, where a Donor dye (ROX) acts as an energy transfer donor to both a FRET Acceptor (DyLt680) and a SET Acceptor (10 ± 0.5 nm AuNP); and Scenario II) a Type I−III RNA Hammerhead ribozyme, where the FRET Acceptor (AF647) is simultaneously an active SET Donor. The Donor dye (Cy3) acts as a donor to both the FRET Acceptor (AF647) and the SET Acceptor (2 ± 0.5 nm AuNP). (B) Schema of contact vectors between labels on the nucleic acid structures. (C) The theoretical SET and FRET efficiency curves for each label pair overlaid with vertical lines indicating the theoretical label−label separation distance. Note that in Scenario I the FRET acceptor is positioned distal to the AuNP in order to avoid energy coupling via SET, while in Scenario II, the FRET Acceptor is positioned so as to couple strongly to AuNP.
and RNA Hammerhead Scenario II constructs (Supporting Figures SF1 and SF2). The lack of a band in the gel at the position corresponding to free DNA or RNA in the assembled McSET constructs supports the high efficiency of coupling to the AuNP following a place exchange reaction. The theoretical contact distances for the appended probes on the two McSET designs are shown in Figure 1B. The distance between the center of the dye and the AuNP metal surface is defined as distance r12, the distance between the pair of fluorescent dyes is distance r23, and the distance from acceptor dye to AuNP surface is distance r13. Models for the DNA/RNA and associated distances (Table 2) were generated from primary sequence and secondary structural folds that were submitted to 3D structure prediction software for estimation of label−label contact distances and compared with available protein data bank (pdb) structures (Material and Methods). In both Scenarios, the theoretical distances in the simulated 3-D structures were calculated using atom positions extracted from PyMol.47,48 In the 3D models, the phosphate-appended C6 thiol linker used to connect the DNA or RNA to the dye and AuNP is accounted for by estimating that the average length of the spacer will be 90% of the all trans-conformation. The predicted dye contact distance in Table 1 was then calculated by assuming a random orientation for the linker relative to the nucleic acid backbone for 10000 simulation events. The corresponding SET and FRET efficiency curves for each McSET design Scenario are plotted in Figure 1C, and the calculated distance of separation for each dye is indicated. The selected AuNP sizes and dye labels in each Scenario were chosen to exhibit roughly equal rates of energy transfer
McSET constructs include a linear B-DNA duplex (Scenario I) and the more complex folded Type I−III Hammerhead RNA ribozyme (Scenario II). As illustrated in Figure 1A, the energy decay pathways for the individual dyes in Scenarios I and II will be directly dependent on the separation distances of the appended probes from the AuNP and by the competition between SET and FRET processes in the McSET constructs. The nucleic acid sequences in Scenario I and II were prepared by commercial synthesis to allow dye and AuNP labels to be incorporated at specified (internal) locations along the nucleic acid backbones. In each case, complementary sequences were annealed via heating in solution followed by snap cooling. The dye-labeled nucleic acid sequences were labeled with a AuNP through a 5′ thiol to form a gold−thiol covalent linkage formed by displacement of Bis(p-sulfonatophenyl) phenylphosphine dihydrate (BSPP) from the surface of synthetically prepared AuNPs by a dynamic exchange process, as described previously.10,14 The loading level of DNA to AuNP is 9:1 in Scenario I and 1:1 for RNA to AuNP in Scenario II. In both cases, the loading levels represent a stochastic distribution of nucleic acids on the individual AuNP surface, and are estimated by UV/vis absorption analysis of AuNP at the LSPR absorption max vs the absorption of the DNA (or RNA) prior to and after cyanide ion etching of the particles in 80 mM NaCN (Supporting Information) following standard methods described previously.10 Visualizations of stained electrophoretic gels of the AuNP− nucleic acid assemblies relative to the unlabeled AuNP clearly exhibit a shift in gel mobility following appendage to the AuNP, confirming the assembly of the McSET DNA duplex Scenario I C
DOI: 10.1021/acsnano.5b05764 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano Table 1. Data from Energy Transfer Measurementsa constants (Å)
lifetimes (ns)
SET coupling and the large separation distance between the dye and the AuNP (Eff = 1.2%).14,16 The ROX donor dye is positioned near the middle of the DNA strand with 19 base pairs (r23 = 72 Å) separation from the DyLt680 and 41 base pairs (r12 = 146 Å) separation from the 10 nm AuNP surface. The placement of ROX along the DNA backbone is near the respective R0 distance for ROX to 10 nm AuNP (SET) and ROX to DyLt680 (FRET); however, the DyLt680 dye is beyond the theoretical distance for SET energy coupling with the 10 nm AuNP when the theoretical atomistic model distance r13 = 215 Å is used for the SET calculation (Supporting Information). The DyLt680 label is appended to the 5′ end, distal from the AuNP, while ROX is labeled within the DNA sequence through an internal base pair modification at a thymine (T) nucleobase. Attachment to the AuNP is through a thiol-terminated 6-carbon spacer linked to the 5′ end of the complementary DNA sequence. The Type II−III Hammerhead ribozyme McSET construct in Scenario II represents a more complex structure and energy transfer pathway scenario. Scenario II explores McSET under conditions where both molecular dyes can energy couple via FRET, as well as both dyes coupling to the 2 nm AuNP via SET. Unlike the duplex DNA structure in Scenario I, the RNA Hammerhead ribozyme is highly flexible due to stem motion about the catalytic core. To model the folded structure accurately, the generated Hammerhead conformation projections in Figure 1A were generated by correlating to a crystal structure for a Mg2+ saturated Hammerhead of similar sequence (PDB ID: 300D).46,50,51,53 The approach was validated in an earlier study where the folding events in a Type I−III Hammerhead RNA ribozyme were monitored using a single color SET experiment.11 As illustrated in Figure 1, in the Hammerhead ribozyme complex, both donor and acceptor dyes couple to the metal nanoparticle through a SET coupling process in addition to the FRET process between the donor and acceptor dye pair. The Hammerhead design thus involves a single FRET pair (Cy3−AF647) and two SET pairs (Cy3−2 nm AuNP and AF647−2 nm AuNP), resulting in competitive relaxation pathways for both dyes appended to the nucleic acid backbone. The 2 nm AuNP is covalently linked to the 5′ end of Stem I and the AF647 dye is linked to the four-nucleotide loop terminating Stem II. The Cy3 donor is attached in arm III, the AF647 dye to the loop on arm II, and the AuNP to stem I. The location of the Cy3 donor dye enables it to report on the proximal location of both Stems I and II simultaneously via SET to the 2 nm AuNP, and via FRET to the AF647 dye label. Thus, AF647 is situated to act as both a FRET acceptor of Cy3 and a SET donor to AuNP. For Scenario II, the dye-dye and dye-particle distances were modeled in saturating Mg2+ conditions to predict distances between labels. The models yielded theoretical separation values between Cy3 and the AuNP (r12) of 44 Å, between Cy3 and AF647 (r23) of 45 Å, and between AF647 and the AuNP (r13) of 35 Å. The experimental design requires more complex modeling of the decay pathway but can report the triangulated global structure by exciting only the donor. Modeling the McSET Energy Pathways. The energy deactivation pathways for the excited state dyes in a McSET construct must be interpreted in terms of the physical properties of the probes (dyes and AuNP) and the separation distance between each probe (r12, r23, and r13) in order to allow a functional mathematical model for McSET to be developed and to allow experimental data to be compared directly with
efficiencies (%)
label
d0
R0
τ0
τ′
τ′Th
EffExp
EffTh
ROX DyLt680 Cy3 AF647
138 71 48 45
67 51 -
4.38 1.68 1.55 1.41
1.78 1.68 0.41 0.37
1.84 1.68 0.34 0.37
0.59 0.74 0.73
0.58 0.78 0.74
a
Experimental lifetimes were measured. Theoretical lifetimes were calculated using theoretical distances and eq 11. Efficiencies were calculated using (Eff = 1 − τ′/τ0). Dashed entries were used when the corresponding data were not applicable; DyLt680 does not act as a donor for SET or FRET, nor does AF647 act as a donor for FRET.
(calculations in Supporting Information). In the calculation, FRET is defined mathematically as coupling energy from the thermally relaxed lowest energy excited state for the donor, and follows a weakly coupled dipole−dipole exchange mechanism for zero-dimensional points in space. Since FRET is modeled at two interacting zero-point dipoles, the efficiency curves in Figure 1C for FRET exhibit an R−6 distance dependence for energy coupling, a value dependent on the energy overlap of the donor and acceptor (⟨J⟩ overlap integral), the molecular orientation of the dyes, and the separation distance.19,49 When a set of dyes, whether engaging in FRET or not, is placed in the near field of the metal, the radiative and nonradiative rates for both the donor and acceptor dye molecules is modified by the metal nanoparticle electric field.14 The efficiency curves in Figure 1C for the SET mechanism are generated by describing the effective enhancement of the radiative and nonradiative rates of the emitting dyes in the near field of the metal nanoparticle, and are described within a strong coupling limit to produce an R−4 energy coupling behavior.14,16 In the case of a gold particle with a radius
