FRET - Förster Resonance Energy Transfer || Multistep FRET and Nanotechnology

13Multistep FRET and NanotechnologyBo Albinsson and Jonas K. Hannestad

13.1Introduction

Excitation energy is transported between molecules in solution, in polymers withchromophoric groups, and in crystalline materials as soon as the sites are closeenough to give an appreciable electronic coupling. If the molecules between whichthe excitation energy transfers are the same species, no change in fluorescenceintensity or lifetime is typically observed. Energy transfer between like molecules,homotransfer, was earlier observed through measurement of the fluorescencedepolarization in solutions with increasing dye concentration [1–3]. Through theF€orster description [4] of singlet excitation energy transfer (fluorescence/F€orsterresonance energy transfer (FRET)) between molecules with very weak coupling,which is typically the case in a solution or between covalently linked chromophoresthat are not connected by a conjugated molecular structure, a quantitative model forthe observed fluorescence depolarization was possible. Although known as aphenomenon but rarely interpreted mechanistically, multistep energy transfer orenergy migration was not discussed in detail until research of the beautifullyarranged chromophore complexes of the natural photosynthesis started. In contrast,normal one-step FRET was quite early realized to have analytical capabilities tomeasure molecular distances difficult to get from other methods [5]. In this chapter,we are going to discuss the phenomenon of multistep FRET together with some ofthe more promising nanotechnology applications based on this process.The use of FRET to measure intermolecular distances and to probe molecular

interactions is nowadays widespread. However, there are instances where singleFRETmeasurement is not sufficient. It might be that the distance separating the twomolecules is too large or that more accurate positioning data than just a distance isrequired. These issues are possible to target by expanding the FRET concept toinclude more than just one donor and one acceptor. Multiple fluorophores can bearranged so that excitation of an initial donor triggers a chain of FRETevents passingthrough multiple chromophores and thus enable energy transfer over longerdistances than is possible with a single donor–acceptor pair. This concept is

FRET – Förster Resonance Energy Transfer: From Theory to Applications, First Edition.Edited by Igor Medintz and Niko Hildebrandt.� 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

j607

most clearly illustrated in the creation of photonic wires, nanoscale assemblieswhere multistep FRET is facilitated by fluorophores arranged in a single file along alinear molecule such as double-stranded DNA. Multichromophoric FRET can alsobe used to obtain more detailed information on the spatial arrangement offluorophores than is provided by the single donor–acceptor pair, which only givesyou the linear distance between them.One area where multistep FRET has been identified to be of particular use is the

growing field of self-assembled nanotechnology. Here, energy transfer could func-tion as an information carrier providing a means to overcome the limitations set bythe wavelength of light.Many of the examples where multistep FRET has been utilized in nanoscale

devices draw their inspiration from the energy migration processes in the light-harvesting complexes in photosynthetic organisms. Here, chromophores arearranged in a way that directs the absorbed sunlight to the reaction center inmultiple steps.

13.2Fundamentals of Multistep FRET

How to describe multistep FRETwill be dictated by the choice of fluorophores. It ishandy to speak of two different regimes: hetero-FRET and homo-FRET. Hetero-FRET follows more or less the same principles as the ordinary two-chromophoricFRETand is the form that is most commonly used. Homo-FRET, on the other hand,is based on energy transfer between identical fluorophore units with sufficientlysmall Stokes shifts. Since the energy transfer occurs between identical molecules,there is no directionality in a homo-FRET system. Furthermore, it is not possibleto selectively excite an individual fluorophore in the pure homo-FRET system.Figure 13.1 shows a schematic energy level diagram and illustrates the principle

Figure 13.1 Schematic energy level diagram showing examples of multistep hetero-FRET (a) andhomo-FRET (b). Arrows indicate one of the many potential energy pathways from the initialabsorption at a donor fluorophore to the emission from the final acceptor.

608j 13 Multistep FRET and Nanotechnology

difference between hetero-FRET and homo-FRET. It is important to note that thefigure shows one possible energy pathway for each case. In a multichromophoricFRET system, multiple energy transfer as well as other de-excitation pathways arealways present. This will be discussed more in detail later in this chapter.

13.2.1Hetero-FRET

In amultichromophoric hetero-FRETsystem consisting of fluorophores A, B, and C,the fluorescence band of fluorophore A overlaps with the absorption band offluorophore B, whose fluorescence band, in turn, overlaps with the absorptionband of fluorophore C. This facilitates the transfer of excitation energy fromA, via B,and further on to C in a cascade of decreasing energy. This cascade can be furtherextended by the adding more fluorophores with absorption and fluorescence oflower energy than C. The limitation is instead practical; with a large number ofdonor–acceptor pairs, the electromagnetic spectrum becomes more and morecrammed. In addition, when the spectral difference between fluorophores is small,it becomes increasingly difficult to avoid cross talk in the system. Because inmultichromophoric FRET it is not only important to minimize direct excitation ofthe acceptor, it is also important to consider energy transfer pathways other than theone going through all the steps. In the simple three-chromophoric FRET (A–B–C),the cases that need consideration are A to C transfer, direct excitation of B (withsubsequent transfer to C), and direct excitation of C (Figure 13.2).The cross talk present in multichromophoric FRET can be utilized to gain

information on the spatial organization of the fluorophores in the studied system.Consider the set of fluorophores A, B, and C, where FRET is possible betweenfluorophores A and C both as A–B–C and directly as A–C. Thus, the resulting output

Figure 13.2 Excitation energy pathways in athree-chromophoric FRET cascade going froma, through b, to c. Radiative and nonradiativedecay from any fluorophore i is represented by

ki and energy transfer between fluorophoresi and j by kij. Direct excitation of fluorophorei is denoted by ei(lex)L(t). Adapted fromRef. [6].

13.2 Fundamentals of Multistep FRET j609

from the system following excitation of A will depend on three separate distances,A to B, B to C, and A to C.By combining these different distances it is possible to obtain an estimate of the

chromophore positions in three dimensions. A detailed description on how todesign two-step FRET measurements was provided by Barkley and coworkers. Inthis model, the integrated excited-state lifetimes of the three fluorophores, followingexcitation of the initial donor, can be described as a function of the energy transferprocesses in the system according to Equation 13.1,

IA ¼ 1= kA þ kAB þ kACð Þ; ð13:1aÞIB ¼ kAB= kA þ kAB þ kACð Þ½ � � 1= kB þ kBCð Þ½ �; ð13:1bÞ

IC ¼ kABkA þ kAB þ kAC

� �kBC

kB þ kBC

� �þ kACkA þ kAB þ kAC

� �1kC

� �; ð13:1cÞ

where Ii, ki, and kij (i, j¼A, B, C) denote the integrated excited-state lifetime inpresence of FRET, the sumof radiative andnonradiative decay fromfluorophore i, andenergy transfer between fluorophores i and j, respectively. To determine the differentenergy transfer rates corresponding to different donor–acceptor distances, multipleintensity measurements are needed. The quenching efficiency for fluorophore Adepends on energy transfer to both fluorophores B and C, as stated in Equation 13.2,

EA;tot ¼ EAB þ EAC ¼ 1� IA=tA; ð13:2Þwhere tA is the excited-state lifetime of A in the absence of FRET. In order todetermine the individual energy transfer efficiencies, it is possible to measure thesensitized emission of fluorophore B in the presence of energy transfer to fluo-rophore C (Equation 13.3),

E0AB ¼ IB=t

0B; ð13:3Þ

where IB is the fluorescence lifetime as described in Equation 13.1b and t0B is thefluorescence lifetime of fluorophore B in presence of the acceptor C. Finally, thesensitized emission from fluorophore C can be described as a function of energytransfer in both one and two steps.

Etot ¼ IC=tC ¼ E0AB � EBC þ E0

AC: ð13:4ÞIn Equation 13.4, Etot denotes the total energy transfer to fluorophore C, boththrough the two-step process (A–B–C) and directly from A to C. CombiningEquations 13.2–13.4, it is possible to extract all interchromophoric distances bymeasuring the quenching of donor A, the sensitized emission of fluorophore B,as well as the sensitized emission of fluorophore C. However, care must betaken to correct for direct excitation of both fluorophores B and C at thewavelength employed to excite fluorophore A. In addition, the same limitationsthat apply to ordinary FRET are of course also valid when determining distanceswith multistep FRET. The measured donor quenching is an average value andthe actual donor separation relies on the dynamics of the studied system. Inaddition, the linking of the fluorophores to the studied system has to be


considered in the same way as for one-step FRET. The ideal case is when thefluorophores are all rigid parts of the studied molecule. However, most often,fluorophores are attached by flexible linkers reducing the precision of distancemeasurements from FRET.

13.2.2Multicolor FRET and Alternating-Laser Excitation

Aside from the conventional methods to detect multichromophoric FRETusing bulksteady-state or time-resolved fluoresce techniques, the vast development of single-molecule fluorescence spectroscopy has led to many new ways of measuringexcitation energy transfer [7]. In order to fully investigate a multichromophoricenergy transfer system, all constituting FRET pairs need to be probed. One of thestrengths inherent in single-molecule techniques is the capacity to probe individualmolecular assemblies one at a time. This makes estimation of the distribution ofdifferent parameters, for example, intermolecular distance, possible. For a multi-FRET system, this means that all the donor–acceptor pairs have to be probed forevery investigated assembly. One way this has been realized is through thetechnique of alternating-laser excitation (ALEX). The term ALEX was coined byShimon Weiss and coworkers and describes a technique based on single-moleculefluorescence microscopy that relies on the combination of multiple, spectrallydistinct, detection channels with a switching between multiple excitation wave-lengths [8]. For a two-color FRET experiment, emission from all constitutingfluorophores is detected simultaneously as the excitation is varied between donoronly and donor and acceptor excitation. The strength in this approach is that it ispossible to distinguish between cases where low FRET efficiency is due to largedonor–acceptor distance and those where it is due to incomplete labeling resultingin donor-only assemblies.Following the initial development of the ALEX technique, its use has been

extended to include more than a single pair of donor and acceptor. This requiresthat the experimental setup has both the capability to switch the excitation wave-length between all the used fluorophores and enough detectors to be able todifferentiate emissions from the different donors from each other. Using ALEXin a multichromophoric FRET system it is possible to probe all the interchromo-phoric distances simultaneously [9]. In addition, the multicolor ALEX approachallows determination of labeling stoichiometry as well as dissection of ensembleheterogeneities.However, as more fluorophores are added, great care must be taken to carefully

select excitation and emission bandwidths to reduce effects from cross talk. Thiscross talk involves leakage of emission into the detection channels of otherfluorophores, direct excitation of acceptors, and energy transfer to multiple, differ-ent, acceptors. In order to accurately characterize the studied system, the extent ofthis cross talk needs to be determined and compensated for. With an increasingnumber of FRET pairs, this correction becomes both more important and exper-imentally complicated.


13.2.3Homo-FRET

In contrast to the cascade in the hetero-FRET case, energy transfer following homo-FRET is not unidirectional. The name homo-FRETcomes from the fact that, insteadof being based on unique donors and acceptors, homo-FRET usually relies on onesingle molecular species. However, The term homo-FRET is a little bit misleadingsince it is not the use of identical fluorophores that is the defining criterion. Rather,the principle relies on fluorophores where the energy difference between theabsorption and fluorescence is very small, allowing energy to go both ways. Thisreasoning can be extended to include more than one fluorophore species given thatit fulfills the stated criterion.In cases when the coupled fluorophores are very closely spaced, energy transfer is

so efficient that it is difficult to talk about individual chromophores. Instead, theexcitation can be considered to be delocalized in the entire system, as a coupledexciton. Such strong coupling generally leads to perturbed electronic states andaltered absorption bands.In complex multichromophoric systems, energy migration can sometimes be

characterized by a combination of strong, delocalized, coupling and through-spacedipole–dipole coupling. Elisabetta Collini and Gregory D. Scholes studied thisintermediate regime using the conjugated polymer poly[2-methoxy,5-(20-ethyl-hexoxy)-1,4-phenylene-vinylene] (MEH-PP) [10]. Energy migration processes werestudied in the polymer systems in two different conformations: one in chloroformpromoting extended chain conformations, and the other in aqueous solution wherean individual polymer chain collapses to form polymer nanoparticles. In thecollapsed nanoparticles, segments of the nanoparticles not directly linked to eachother through covalent bonds are close together in space, compared with theextended conformations where only the neighboring polymer segments are adja-cent. From this, it is likely that the dense nanoparticles will display substantialthrough-space interchain energy migration, whereas energy migration in theextended polymer chain in chloroform will predominantly be of the coherent,strongly coupled, intrachain type. In order to measure the two different migratoryprocesses, two-time anisotropy decay (TTAD) experiments were performed. Theanisotropy decay is recorded as a function of two time delays, t and T. The time delay,T, is the population time, during which excited-state dynamics, such as excitationenergy transfer, occurs. The second time lag, t, scans a time period when the systemis in coherence between ground and excited electronic states. The experiments showthat the anisotropy decays more rapidly during T for the nanoparticles, than for theextended polymer chain. This can be explained by the interchain energy transfer,made possible by the compact structure of the nanoparticles. In contrast, only theextended MEH-PPV in chloroform showed anisotropy decay during t. Thus, in acomplex system, different chromophore conformations can lead to differences inenergy migration mechanisms.Calculating the energy transfer efficiencies in a homo-FRET system is somewhat

different compared with the hetero-FRET case because of the multidirectionality of


the energy transfer. Figure 13.3 shows a homo-FRET system consisting of fouridentical fluorophores. Because all fluorophores have the same spectral properties, itis not possible to selectively excite individual fluorophores in a pure homo-FRETsystem.The energy transfer scheme presented in Figure 13.3 is not limited to the case

with just four fluorophores. Equation 13.5 describes the change in excitation energyat any given fluorophore, n, in a general homo-FRETsystemwith the total number ofm fluorophores.

dIndt

¼ en lexð ÞL tð Þ þXmi 6¼n

kinIi � kn þ knið ÞInð Þ ð13:5Þ

As can be seen from Equation 13.5, in a true homo-FRET system, the energytransfer has no directionality, as the rate is equal in both directions between each“donor” and “acceptor.” It is also not possible to investigate homo-FRETefficienciesusing, for example, donor quenching, since both donors and acceptors are spectro-scopically identical. However, information from fluorescence anisotropy can beutilized to gain information on homo-FRET processes in multichromophoricsystems. As FRET has a depolarizing effect on the emission, the degree ofanisotropy can be used to measure the extent of energy transfer in a system, which,in turn, is related to the number of participating fluorophores within energy transferrange, and thus a measure for the compactness of the multichromophoric system.Kaminski and coworkers show in a review article how homo-FRET can be used inconjunction with fluorescence anisotropy imaging (FAIM) to study assembly andaggregation of macro- or supramolecular structures [11]. Examples where thistechnique has been put to use include studies on amyloid aggregation [12], lipidrafts [13–15], and protein clusters [16,17].One fluorophore with substantial capability for homo-FRET is the DNA-binding

fluorophore YO [18–20]. YO is virtually nonfluorescent in solution, but in thepresence of DNA, the fluorescence quantum yield increases by about a 1000-fold.YO has been shown to bind to DNA through intercalation between the base pairs.This is true for fluorophore/base pair ratios up to 0.2. At higher YO concentrations,

Figure 13.3 A homo-FRET system consisting of four identical fluorophores with indicatedabsorption, emission, and energy transfer pathways. Because of the homo-FRET conditions, allpossible energy transfer pathways are bidirectional.


other modes of binding become apparent [20]. An effort to quantify the fluorescencedepolarization caused by homo-FRET in a system with noncovalent DNA binders,such as YO, PO, and DAPI, wasmade by Albinsson and coworkers [19]. In this work,fluorescence depolarization was described by a semiempirical Markov chain model.In this model, the degree of fluorescence depolarization is investigated as a functionof the intercalator density. The depolarization caused by energy transfer is depen-dent on the relative orientation of the donor and acceptor emission and absorptiontransition dipoles, respectively. This is, in turn, determined by the relative positionsof the donor and acceptor in the DNA strand (i.e., the number of base pairsseparating them). For a given random distribution of intercalated YO molecules,there is a set of energy transfer and emission probabilities for each fluorophore inthe system. These probabilities define the energy migration pathway and thus thefluorescence depolarization. The simulation algorithm is constructed in steps whereexcitation energy located at any of the fluorophores can either be emitted ortransferred to any of the other fluorophores in the system. In this way, thedistribution of excitation energy at any given step, vn, is defined as a product ofthe excitation energy distribution in the previous step, vn�1, and the complete setof emission and energy transfer probabilities, M. By repeating this recursivecalculation, it is possible to express the excitation energy distribution at anypoint as a function of the initial excitation energy distribution, v0, according toEquation 13.6

vn ¼ vn�1 �M ¼ . . . ¼ v0 �Mn: ð13:6ÞThe steady-state product, v1, is obtained by letting the number of steps approach

infinity, as shown in Equation 13.7

v1 ¼ v0 limn!1Mn: ð13:7Þ

Thus, the resulting depolarization is obtained as a product of all the energytransfer steps leading up to the eventual emission of a photon and the removal of theexcitation energy from the system. The simulations show an expected decrease inemission polarization with increasing dye concentration. The steepness of thedecrease is dependent on the F€orster radius for homo-FRET between the interca-lated dyes. Figure 13.4 shows simulation outcomes for different set values of R0,

ranging between 10 and 60A�. At 10A

�, there is a linear dependence of the degree of

depolarization on the dye-to-base pair ratio. For FRET pairs with larger F€orster

radius, R0� 29A�, the expected limiting depolarization value, 0.25 or r¼ 0.1, is

reached around 0.2 intercalators per base pair. This model can be extended toaccommodate any dye bound to any linear scaffoldingmolecule. However, caremustbe taken to accurately describe both the angle between the dye transition momentand the long axis of the scaffold molecule, as well as the relative orientation of thedyes as a function of their linear separation. The usefulness of the model will beexplored later on in the chapter, where systems with more than one type offluorophore, and cases with a distinct directionality in the energy transfer, arediscussed.


13.3Energy Transfer in Photosynthesis

Photosynthetic organisms have developed an intricate set of chromophoric arrange-ments to gather energy from sunlight. Here we will exemplify this by describing howphotosynthetic bacteria use a remarkable example of nanoscale engineering to feedthe reaction center with excitation energy. This text should not be read as a completereview of the function of the bacterial light-harvesting complexes. Instead, we wantto provide a brief overview, largely because the light-harvesting complexes have hadan important role to inspire many of the nanoscale photonic assemblies that wepresent in the chapter. This chapter is, to a large extent, based on the work by KlausSchulten and coworkers [21–23].In bacterial photosynthesis, a closely spaced pair of bacteriochlorophylls (BChls),

called the special pair, is responsible for an excited-state electron transfer reactionthat leads to polarization of the cell membrane. Instead of letting the sunlight justexcite the special pair in the reaction center directly, photosynthetic bacteria arrangea large number of circular light-harvesting complexes around the reaction centerthat channels the excitation energy into it in multiple steps. The pigment molecules,mostly bacteriochlorophylls, responsible for channeling the excitation energy to thespecial pair vastly outnumber the reaction center complexes; often the ratio is on theorder of hundreds to thousands pigment molecules per reaction center.Why do the photosynthetic bacteria organize the gathering of light energy by

having a large number of light-harvesting protein complexes delivering excitationenergy to a small number of reaction centers (RC)s instead of just having morereaction centers? There are multiple reasons for this. One primary reason for this isthe turnover rate of the reaction center. The reaction center is capable of undergoingthe electron transfer reaction at a rate of 1000Hz. However, the chlorophylls of the

Figure 13.4 Computer simulated degree of depolarization values as function of dye to base pairratio (D/B) for different homo-FRET F€orster distances. The curves represent, from top to bottom,R0 values of 10, 13, 17, 29, 37, and 60A

�. Adapted from Ref. [19].

13.3 Energy Transfer in Photosynthesis j615

RC that feeds the reaction with energy absorb only at 10Hz in direct sunlight and0.1Hz in dim light. Thus, the reaction center is not capable of functioning atmaximum turnover on its own. Instead, the light-harvesting complexes help thereaction center to reach maximum turnover by feeding it with excitation energy.Another reason is that the light-harvesting complexes absorb light at differentwavelengths than the reaction center, making it possible for the bacterium to makemore use out of the sunlight. In addition, the light-harvesting complexes are smallerand less complex than the reaction center and therefore require less energy from theorganism’s metabolic system to synthesize.In the bacterial photosynthesis, there are two different ring-shaped protein

complexes responsible for directing excitation energy to the reaction center, namedLH-I and LH-II (Figure 13.5). LH-I consists of 16 identical building blocks calledab-heterodimers arranged in a circle around the reaction center. Theab-heterodimer is a complex that consists of one a-apoprotein, one b-apoprotein,three BChls and one carotenoid. The a-apoprotein and the b-apoprotein are botha-helices of 56 and 45 amino acids, respectively. The constituent responsible forchanneling the excitation energy to the reaction center is a ring-shape assemblage ofthe 48 bacteriochlorophylls. This ring, LH-I, is formed around the reaction center towhich LH-I is strongly associated.The LH-II ring is slightly smaller than LH-I, about 14 nm in diameter, and is

positioned in multiple copies in between the RC–LH-I assemblies. LH-II consists ofeight ab-heterodimers, compared with the 16 of LH-I. The proteins scaffold24 BChls and eight carotenoids. Sixteen of the BChls form a ring-shaped aggregateknown as B850. The remaining eight BChls form a second ring that is looser. Thisring is denoted B800. The eight carotenoids span the distance between the BChls,each touching two separate bacteriochlorophylls.The energy transfer cascade starts in the LH-II complexes. LH-II has the ability to

absorb light at both 850 nm through the bacteriochlorophyll ring and at 800 nmthough the carotenoid-associated bacteriochlorophylls. Excitation energy absorbedby the B800 BChls can be transferred to the B850 by FRET. In the B850 ring, photons

Figure 13.5 Schematic image of the bacteriallight-harvesting system with the reaction centertogether with the surrounding LH-I and anumber of LH-II assemblies positioned outside

LH-I. The white arrows indicate the flow ofexcitation energy passing through LH-II andLH-I before reaching the RC. Adapted fromRef. [22].


are absorbed by only two of the 16 possible exciton states, the second and thirdenergetically lowest. As the lowest excited state is optically forbidden, the B850 ringacts as an energy storage unit, preserving excitation energy until it can be forwardedfurther in the cascade toward the RC. LH-I absorbs at lower energies than LH-II(875 nm). This makes it possible to funnel excitation energy from multiple LH-IIthrough LH-I to the RC. The photoactive component in all these assemblies is theBChl-a unit. Since the absorptivity of monomeric BChl-a peaks at 772 nm, the 800and 850 nm absorption of the two LH-II rings as well as the 875 nm absorption ofLH-I are results of the spatial organization of the fluorophores. Suggested explan-ations for this are excitonic interactions as well as BChl–protein interactions. Theenergy transfer from LH-II, via LH-I, to RC occurs within 100 ps at 95% efficiency.Besides the high efficiency and the high rate, one remarkable feature is the numberof potential pathways leading to the reaction center. One possible pathway, usingexcitation at 800 nm, involves four sequential steps. Starting with initial excitation ofB800, excitation energy is transferred to the B850 ring in the same LH-II assembly.Excitation energy is then, in turn, transferred to the B850 ring of the next LH-II. Thisis the followed by energy transfer to LH-I, which finally transfers the excitationenergy to the reaction center.The fascinating features of the bacterial light-harvesting systems show how spatial

organization and precise tuning of absorption and emission properties can be usedto construct systems for efficient excitation energy transfer with high spatialprecision. How these principles can be applied in nanoscale devices will bediscussed in the next section.

13.4Photonic Wires and Multistep FRET in Nanotechnology

The development of bottom–up design strategies in nanotechnology has sparked adesire for new means of information transfer on the nanometer scale. As theresolution of conventional optical techniques is limited by the diffraction of light,their usefulness is limited when the relevant length scales are in the order of tens ofnanometers. Multistep FRET combines the precision from conventional FRETwitha reach that extends to more than 10 nm, making an interesting candidate fornanoscale communication. This device-oriented approach is a fairly new view onFRETwhere it is used not only as a reporter for interaction, but also as a function inits own right.

13.4.1Photonic Wires

One of the most basic examples of the use of multistep FRET in nanotechnology isthe creation of so-called photonic wires. This could be compared with an opticalwaveguide where light energy is transported linearly from one end to the other. In

13.4 Photonic Wires and Multistep FRET in Nanotechnology j617

the case of the FRET-based photonic wires, fluorophores are arranged in a linearfashion to achieve directional energy transfer from one end to the other. The wirecan either be constructed directly out of the coupled fluorophores or some kind ofmolecular scaffold can be used to arrange the fluorophore in a wire-like fashion.Perhaps the most prominent example of this is arrays of linked porphyrinmolecules[24–26]. The first example of the concept of a molecular photonic wire, drawinginspiration from natural light-harvesting complexes was presented by Lindsey andWagner in 1994 [25]. In this work, a supramolecular assembly consisting of a boron–dipyrromethene (BODIPY) input unit, three zinc porphyrins, and a free-base (FB)porphyrin covalently connected through diarylethyne linkers was presented(Figure 13.6).Emission was predominately observed from the free-base porphyrin following

excitation of BODIPY, showing that excitation energy had been transferred along theconstructed wire. The energy flow in the porphyrin-based wire was shown to begoverned by a combination of Dexter-type through bond superexchange-mediatedinteractions and FRET [27]. Porphyrins are especially interesting because of theirresemblance to the chlorophylls important in, for example, the natural light-harvesting complexes previously discussed.One of the simplest examples of what could be considered as a photonic wire is the

perylene trimer studied by Hernando et al., which consists of three identical tetra-phenoxy-perylene diimide chromophoric units covalently linked together [28]. Therigid structure of the trimers holds the chromophoric units locked in a perpendic-ular conformation, thus preventing full extension of the conjugated p-system. This,in combination with a short interchromophoric distance of only 1.3 nm, results instrong exciton coupling. This results in a delocalization of the excitation over theentire multichromophoric system. The strong coupling enables propagation ofexcitation energy in the multichromophoric system without the presence of a strongdriving force in the form of a steep energetic gradient. With the addition of a high-energy dye at one extreme of the wire and a low energy dye at the other, a highlyefficient end-to-end transfer could be achieved. However, because of the shortinterchromophoric distance in the perylene arrays, their usefulness as wires innanoscale systems is fairly limited. Another type of system having strongdelocalization of excitation energy is linear and circular arrays of conjugatedporphyrins.An alternative approach is to use some kind of scaffolding molecule, for example,

DNA, for arranging the fluorophores. The fluorophores have to be selected so that

Figure 13.6 Supramolecular photonic wire consisting of an input dye, BODIPY (blue), three Znporphyrins (green), and an output dye, free-base porphyrin (yellow). Adapted from Ref. [25].


there is an energy gradient going from one end to the other. Thismakes it possible toexcite the edge fluorophore in the high-energy end and obtain multistep energytransfer all the way to the other end. An early example of multistep FRETcoupled toDNA comes from the work of Nicholas J. Turro and coworkers [29]. Here, threefluorophores, 6-carboxyfluorescein (F), N,N,N0,N0-tetramethyl-6-carboxyrhodamine(R), and Cy5 (C) were coupled to a 26 base pairs long, single-stranded DNAmolecule(Figure 13.7).This multichromophoric assembly is not designed to function as a wire, convey-

ing excitation energy between two spatially separate points. Rather, the fluorophoresfunction together to form a nanoscale dye with a Stokes shift of 182 nm. Thequenching of the initial donor, F, is 99% and the assembly has an overall fluores-cence quantum yield of 0.13. Even before this, Uchimaru and coworkers hadintroduced a sequential arrangement of donor and acceptor dyes in double-strandedDNA [30]. The fluorophore 6-carboxyfluorescein (F) was coupled to a 25-mer scaffoldstrand. This fluorophore acts as the initial donor in a three-chromophoric system.The other two fluorophores, 4,7,20,40,50,70-hexachloro-6-carboxyfluorescein (H) and6-carboxy-X-rhodamine (R) are coupled to a 15-mer and a 10-mer oligonucleotide,respectively, both of which bind to the 25-mer scaffold. The total end-to-end distance

is approximately 80A�, which is within the range that can be covered by single-step

FRET. The assembly is not labeled as a “photonic wire,” instead the intermediatestep is used as an enhancer of the FRET process between the fluorophores at eachend of the DNA strand. The intermediate fluorophores were shown to enhanceFRET between the end fluorophores by 360%. These results provide a basis for theuse of DNA-based FRET over even longer distances, something that came to bereferred to as molecular photonic wires or DNA-based photonic wires. The conceptof the DNA-based photonic wire was first introduced by Sauer and coworkers [31].Here, DNA was used a scaffold for a series of fluorophores creating a photonic wirebased on the cascade principle [31,32]. In conjunction to this, it is also appropriate tomention the work by Ohya et al. [33]. Although, the term “wire” is not mentioned inthe paper, the work carries many of the characteristics associated with the term.Yuichi Ohya’s take on the DNA-based molecular wire is discussed more in depthlater in the chapter.Because of the increase in the number of high-brightness (highmolar absorptivity

and high quantum yield) fluorophores for nucleic acid labeling, there are a largenumber of donor–acceptor combinations available that can be used in a multistepenergy transfer cascade. In the work by Sauer and coworkers, a 60-mer single-stranded DNA molecule is used as a template strand [31,32,34,35]. To this strand,complementary 20-mer DNA strands are hybridized. The 20 base pair oligomers arelabeled with fluorophores which, when all strands hybridize with the template, forma chain of energy transfer donor–acceptors (Figure 13.8a). Here, the sequencespecificity of the base pairing of the labeled strands with the template strand is usedto assure that the fluorophores arrange in the correct order to achieve end-to-endenergy transfer in the wire. The fluorophores are placed so that, when all strandsassemble, there is a 10 base pair separation between them. The choice to use a 10


O

O

PO O–

O

HO

N

HN

O

O

NH

N

OH

O

C

O OHO

O

O–

O

O

PO O–

O

N

HN

O

O

CH3

O

O

PO O–

O

N

HN

O

O

NH

N

OH

O

C

O N+N

O

O–

4

O

O

PO O–

O

N

HN

O

O

CH3

6 O

O

PO O–

O

N

HN

O

O

NH

O

NH

N

O

–O3S

N+

SO3–

O

O

PO O–

O

N

HN

O

O

CH3

12 O

N

N

O

NH2

OH

F

Cy

R

ET-1ET-2

ET-3

Excitation488 nm Emission

525 nm

Emission585 nm

Emission670 nm

0.30

0.23 Excitation 488 nm

Emission 670 nm

Abs

orba

nce

0.15

0.08

400 500 600

Wavelength (nm)

700 800

1.00

0.75

Fluorescence intensity (a.u.)

0.50

0.25

Figure 13.7 DNA-linked three-chromophoric FRET system together with absorption (dashedline) and emission (solid line) spectra. Adapted from Ref. [29].


base pair separation between the fluorophores in the FRETcascade is not by chance;as 10 base pairs correspond to one turn of the B-DNA helix, all fluorophores will bepositioned on the same side of the DNA duplex. This increases the potentialefficiency of the wire since all donor–acceptor pairs are positioned along the axisof the wire instead of having a parallel and a perpendicular component, which wouldbe the case if the fluorophores were positioned on opposite sides of the DNA duplex.In addition, a 3.4 nm separation (corresponding to 10 base pairs) between thefluorophores in the cascade will result in one-step FRETefficiencies of 91–98.7% for

donor–acceptor pairs with F€orster distances of 50 and 70A�, respectively. Thus, these

fluorophores provide a potential for high overall wire efficiency. However, multiplestudies on photonic wires of this type have shown lower experimental end-to-endenergy transfer efficiencies than what could be expected from the F€orster distancesof the individual donor–acceptor pairs [34,35]. The same principal design has beenused for a number of wires, comprising different fluorophores. In one wire, thefluorophore combination Rhodamine green (RhG), Tetramethylrhodamine (TMR),ATTO 590, LightCycler Red (LCR), and ATTO 680 was used [31,32]. In the secondwire, LCR was replaced by ATTO 620 [34,35]. In both cases, the fluorophores wereplaced with a 10 base pair separation, which would correspond to individual FRETefficiencies above 90%. This should have, in turn, yielded an overall efficiency of atleast 70%. Instead, the observed overall efficiency was of the order of 15–30%However, single-moleculemeasurements revealed subpopulations of wires in whicharound 90% of the emission was detected in the channel corresponding to thelowest-energy dye. End-to-end energy migration was further studied by sequentiallybleaching the wires that emitted in the red channel. These experiments showed thatthe terminal acceptor is bleached first, followed by the higher energy dyes in order,with the initial donor bleaching last. The emission intensity profile also revealed that

Figure 13.8 (a) DNA-based photonic wirecontaining five different fluorophores formingan energy transfer cascade that spans from oneend to the other. Adapted from Ref. [32].(b) False colored fluorescence micrographs of

single photonic wires with increasing lengthand number of fluorophores (1–5). The resultsshow that not all wires emit from the mostredshifted dye. Adapted from Ref. [34].


emission from the wires were, at any time, dominated by a single fluorophore, thatis, energy transfer is efficient up to the last available dye in the cascade [31].The observed heterogeneity of the photonic wires was further studied using

single-molecule fluorescence spectroscopy [34,35]. Although all wire componentshave been added to the sample, there is a large fraction of the wires that showspredominant emission from one of the linking fluorophores (Figure 13.8b). Theprevalent explanation for this is focused on the assembly of the DNA duplex.Because of the arrangement with the different dyes covalently linked to shortoligomers that bind to a longer template strand, the assembly of the wires requiresnot only for two single-stranded DNA molecules to form a duplex, but for several ofthese events to occur. This places a higher requirement on strand specificity andcontrol of, for example, hairpin formation than if two strands of equal length wereused. Other observations point to chromophore orientation and molecular environ-ment as factors that affect the energy transfer process in a negative way.In the photonic wire, there are two principle loss terms. First, the directionality

implies a decrease in excitation energy by going from one end of the wire to theother. This redshift means that there is less energy to use for driving photochemicalreactions at the output of the wire. Second, excitation energy can be lost due tophotophysical processes, for example, fluorescence and internal conversion, otherthan FRET; that is, the wire can be said to have a specific quantum yield for end-to-end transfer that depends on the number of fluorophores and on how well theycouple. These two parameters have to be balanced against each other in the design ofthe wire system. The redshift is minimized by limiting the number of differentfluorophores in the system, since every new donor–acceptor pair leads to lowerexcitation energy. However, a high quantum yield is obtained by having strongcouplings between the fluorophores, that is, short donor–acceptor distances. But,short distances means that we need many donor–acceptor pairs to extend the wire,thus increasing the redshift. Carefully choosing the fluorophores and their relativepositions can strongly influence the efficiency of the wire. For instance, it isunnecessary to place the donor and the acceptor closer than roughly 50% of R0

because that corresponds to more than 98% efficiency and there is only very little togain by moving the fluorophores closer.The redshift problem can also be tackled in another way. Instead of building the

wire through an energy transfer cascade going from one end to the other, the middlepart of the wire can be constructed from a sequence of fluorophores capable ofhomo-FRET. In this mediating homo-FRET part of the wire, the energy transferproceeds in a way that can be described by one-dimensional statistical diffusion. Ifthe fluorophores are separated by an equal distance, there is no net driving force forenergy transfer in either direction of the wire. Instead, directionality is achieved byflanking the homo-FRETdyes with a higher energy dye (injector) on one side and alower energy dye (detector) on the other side. In this way, it is possible to direct theflow of excitation energy without losing too much energy in energy transfercascades.Already in 2003, Ohya et al. presented a photonic wire where homo-FRET was

used to achieve end-to-end energy transfer. In this work, a template oligomer (20, 30,


or 40-mer) is combined with shorter, dye-labeled, 10-mer strands to generate thephotonic wire [33]. The wire consists of the three fluorophores, eosin (Eo), Texas Red(TR), and tetramethylrhodamine (Rho), each covalently attached to a 10-mer DNA,complementary to the longer template strand. The strands assemble in an order thatcreates an energy transfer cascade going from one end of the wire to the other. Forthe 30- and 40-mer wires, the Rho unit is present in one and two copies, respectively.When there are two Rho units present in the wire, as is the case for the 40-mer wire,part of the energy transfer is mediated through homo-FRET between the identicalRho units. Fluorescence measurements showed no energy transfer between eosinand Texas Red in the absence of tetramethylrhodamine, both for the 30-mer and forthe 40-mer. When DNA strands with Rho attached in the 50 end were added to thesystem, a 23.7 and 21.6% energy transfer efficiency was detected for the 30-mer andfor the 40-mer, respectively. Ohya et al. used the formula presented in Equation 13.8to calculate the energy transfer efficiency,

Tapp %ð Þ ¼ A Xð Þ=A 100ð Þ � 100 ¼ wDFD=wAFA 1� Tð Þ þ Tf g � 100; ð13:8Þ

where Tapp is the apparent energy transfer efficiency and A(100) and A(x) are theintegrated emission and excitation spectra, respectively. The terms w and F repre-sent the quantum yield and fluorescence intensity, respectively, for both donor (D)and acceptor (A) fluorophores. T is the actual energy transfer efficiency. Theequation is adapted from Imanishi and coworkers [36] and compensates for thespectral overlap between donor and acceptor emissions. This model does notaccount for the fraction of the acceptor emission that is due to direct excitationof the mediator dyes rather than the original donor, something that is relevant in thecase of multichromophoric FRET. In many systems, this effect is pronounced, as wewill see later on.Quake and coworkers presented another example of a DNA-based photonic wire

utilizing homo-FRET to facilitate end-to-end energy transfer. The photonic wire isbuilt using the fluorophores 6-carboxyfluorescein (6-FAM), 6-tetramethylrhod-amine-5(6)-carboxamide (TAMRA), and Cy5 [37]. In this work, fluorophores arecovalently attached to either one of the two equally long DNA strands forming thefinal duplex. The fluorophores are positioned with a separation of 10 base pairsputting them either on the same side or on opposite sides of the duplex dependingon whether they are attached to the same or to different DNA strands, respectively.The fluorophore TAMRA functions as a mediating unit connecting 6-FAM and Cy5and the number of TAMRA units depends on the length of the wire. For the wirebased on a 20-mer DNA strand, only one TAMRA fluorophore is used to bridge thegap between injector and detector dyes. Since only one mediating unit is present inthe assembled wire, there is no homo-FREToccurring. Instead, end-to-end transferoccurs primarily through the two-step 6-FAM–TAMRA–Cy5 FRET (there is also aslight contribution of direct 6-FAM to Cy3 FRET to the end-to-end transfer)(Figure 13.9b). An extension of the wire to a 40-mer DNA means inclusion oftwo more TAMRA units to a total of three. Here, substantial homo-FRET betweenthe mediating TAMRA fluorophores is required for energy to be transferred from


the injector to the detector end of the wire (Figure 13.9b). The amount of end-to-endtransfer in the longer wire is dependent on the number of mediating units. Removalof the middle of the three TAMRAunits results in a decrease in the Cy5 fluorescencetogether with an increase in the 6-FAM fluorescence. However, there is stillsuccessful end-to-end energy transfer in the system.So far, we have considered only covalently linked fluorophores whose positions are

well defined and controllable. However, photonic wires can also be created usingfluorophores that interact noncovalently with a wire scaffold. For a DNA-based wire,there are two principal ways in which a fluorophore can associate with the DNAstrand to assemble the wire, either through intercalation or by binding to the majoror minor groove. In these cases where the fluorophores are not covalently bound tospecific positions in the DNA strand, there is no control over fluorophore distribu-tion, it is therefore only possible to create a photonic wire based on this approachwith fluorophores capable of homo-FRET. In the DNA case, such dyes are, forexample, the cyanines YO-PRO or TO-PRO.A DNA-based photonic wire built using the dyes Pacific Blue, YO-PRO-1, and Cy3

was constructed by Albinsson and coworkers [38]. The intercalator YO-PRO medi-ates energy transfer from Pacific Blue to Cy3 by homo-FRET. A schematic imageof the assembled wire is presented in Figure 13.10a. Because the intercalation of

Figure 13.9 (a) Emission spectra from different 20-mer wire constructs together with extractedTAMRA emission from wires with and without Cy5. (b) Emission spectra from different 40-merwire constructs together with extracted Cy5 emission. Adapted from Ref. [37].


YO-PRO is not sequence specific, there is no control of the positioning of themediating units. However, increasing the mediator concentration leads to higheremission from the output dye Cy3 (Figure 13.10b).As with the photonic wire created by Quake and coworkers [37], the homo-FRET

nature of the wire depends on the length of the DNA strands. However, as themediator inserts randomly into the DNA strand, there is no controllable way toobtain wires showing strict two-step FRET. Instead, a short wire (i.e., 20-mer) willpredominately show a two-step FRET from Pacific Blue, via a single YO-PRO to Cy3at low YO loading ratios.In this kind of wire, the end-to-end efficiency as a function of intercalator density

follows approximately a sigmoidal function (Figure 13.10c). For shorter wires, theinitial lag phase is nonexistent, as end-to-end energy transfer exists to some extenteven without mediators. As intercalators are added to the system, the end-to-endenergy transfer efficiency immediately starts to increase to finally level out at highermediator densities. For a 20-mer wire, this plateau level is reached at approximately0.15 YO-PRO/base pair giving a 90% end-to-end energy transfer efficiency. Ideally,the maximum value for end-to-end transfer efficiency would be 1. However, thiswould require a high loading of mediators with almost perfect delocalization of theexcitation across the whole wire. Assuming nearest neighbor exclusion, there is a

Figure 13.10 (a) Self-assembled DNA-basedphotonic wire consisting of the fluorophoresPacific Blue, YO-PRO, and Cy3. YO-PRO acts asa mediator conveying the excitation energy fromPacific Blue to Cy3 through diffusive homo-FRET. (b) Emission spectra for an assembled

20-mer wire with varying YO concentration. Asthe amount of YO increases, the Pacific Blueemission is quenched and the Cy3 emission isenhanced. (c) Experimental and simulated end-to-end efficiencies for 20-mer and 50-mer wires.Adapted from Ref. [38].


finite limit to how densely the mediating units can be packed in the DNA-basedphotonic wires. In a short wire, this might be sufficient to obtain end-to-end transferof excitation energy with 100% efficiency. However, as more and more intercalatordyes are added to the system, the influence of alternative binding modes, other thanintercalation, becomes influential [19]. With higher mediator loading there is also anincreased probability that two intercalations that are in close proximity formenergetic sinks that hinder the propagation of excitation energy. These trap statesprevent a 100% efficiency from being practically obtainable. For a longer wire (e.g.,50-mer), a low number of mediators is not sufficient to achieve connection betweeninput and output dyes. Therefore, an initial lag phase is observed with no increase inend-to-end transfer efficiency with increasingmediator density. In the case of the 50-mer wire, the connection between input and output of the wire is not strong enoughto obtain a limit where the increase in transfer efficiency levels off with increasingmediator density, even when the intercalator concentration approaches the maxi-mum value of 0.5 YO-PRO/base pair. For the 50-mer wire, this plateau is obtained atapproximately 30% end-to-end energy transfer efficiency.The noncovalent approach brings up other important aspects that need to be

considered in the case of nanoscale technology: self-assembly and bottom-up design.Noncovalent association of the constituting fluorophores means that the creation ofthe wire requires fewer controlled linkage steps forcing the fluorophores into theircorrect positions. For more complex systems, extensive modifications are alimitation.In an effort to tackle the uncertainties associated with the sequence nonspecific

binding of YO-PRO to DNA and to create a more programmable photonic wire,Burley and coworkers have constructed system where the YO-PRO mediator bindsto a specific base pair sequence. This is achieved by tethering the fluorophore topyrrole–imidazole polyamides (PAs) (Figure 13.11a) [39]. The polyamides were

Figure 13.11 (a) Polyamide used for sequence specific binding to DNA. (b) 21-mer wire showingthe binding motif in bold. (c) 55-mer wire with four binding sites. Adapted from Ref. [39].


initially constructed by the Dervan group and bound to the minor groove specificallytargeting the sequence 50-WWGGWCW-30, where W is either adenine or thymine[40,41]. Through the specific binding of the PAs to DNA, the intercalation of YO-PRO can be directed to a certain binding site (Figure 13.11b).A trade-off of using this method to obtain sequence specificity is that it involves a

reduction in the possible mediator density. As the binding motif for the polyamidemoiety responsible for the sequence specificity is larger than the base pair occupancyof one YO-PRO molecule, it is not possible to pack the intercalators as closely as ifthey were binding randomly to the DNA strand. Thus, the major advantage of thesequence-specific binding mainly appears at low mediator concentrations when it ispossible to achieve a much more favorable intercalator distribution by controllingtheir positions in the wire. The different wire lengths that were created were one 21-mer (Figure 13.11b), one 55-mer (Figure 13.11c), and one 80-mer having one, four,and six binding sites, respectively. An asymmetric core polyamide, recognizing thesequence 50-ATGGACA-30, was selected, both for its high binding affinity to DNA aswell as for its directional binding. The YO intercalator is attached to the PA by atetherer with a length corresponding to two base pairs. This means that the totalbinding site size is nine base pairs, in addition to the space occupied by theintercalator itself. The photonic wire was assembled using Pacific Blue as donor, YO-PA asmediator, and Cy3 as acceptor, similar to the work by Albinsson and coworkersdescribed above [38]. The effect of the sequence-specific mediator binding wasinvestigated by comparing the enhanced emission of Cy3 on Pacific Blue excitationwith YO added to the system, both for wires with and without sequence-specificmediator binding. Adding one equivalent of the sequence-specific YO assembly tothe 21-mer DNAwire results in a threefold increase of Cy3 emission, compared withwhen a randomly binding YO assembly is added. Because the emission intensitiesfrom Pacific Blue and YO were comparable in both cases, the Cy3 enhancement isattributed to increased efficiency for the transfer of excitation energy from YO toCy3. The increase in energy transfer efficiency for the 21-mer wire that is associatedwith directed mediator binding is not surprising. Since a distance of 10–11 basepairs is well within single-step FRET distance, there is only need for a single,correctly placed, mediator molecule to enable the two-step end-to-end energytransfer (i.e., Pacific Blue–YO–Cy3). When mediators bind randomly, many, lessefficient, conformations are enabled. Thus, a higher loading is required when themediators bind randomly to achieve efficient end-to-end transfer. Increasing theconcentration of randomly binding YO to three equivalents resulted in Cy3 emissionlevels comparable with those of the sample with sequence-specific binding. How-ever, this also resulted in a fourfold enhancement of the YO emission, indicatingthat many of the YO molecules bind in positions not favoring end-to-end FRET. Forthe longer wire lengths (55-mer and 80-mer), there is a substantial influence of YO toYO homo-FRET in the end-to-end energy transfer process. For the 55-mer, anintercalator concentration of four equivalents (i.e., one intercalator per binding site)results in a twofold increase in the Cy3 emission for the sequence-specific YO–PAsample compared with the sample with nonspecific binding of YO. Compared withthe 21-mer, a substantial amount of excitation energy is trapped in the YO part of the


55-mer wire, both for the sample with and without sequence-specific binding of YO.This effect is reflected in the end-to-end efficiency of the YO–PA wires, which is 49and 26% for the 21-mer and the 55-mer wires, respectively. For 80-mer wire (with sixequivalents of YO–PA), the enhancement decreases further to 1.5 and the resultingend-to-end efficiency is 14%. Thus, the advantage of sequence-specific bindingdecreases with increasing wire length. Instead, the footprint of the polyamide unitbecomes disadvantageous since it prevents dense packing of mediator dyes.However, it is clear that programmable binding of mediator molecules substantiallyenhances energy migration making it possible, even at relatively low bindingdensities, to achieve end-to-end transfer of excitation energy over a 27 nm distance(for the 80-mer wire).Besides sequence-specific binding, there is another way that the arrangement of

the communicating fluorophores can be controlled to achieve higher FRET effi-ciency. The relative orientation of donor and acceptor dyes strongly influences theenergy transfer efficiency. In DNA, the orientation will depend on the turning of theDNA helix, and thus on the number of base pairs separating the donor and acceptormolecules. The effect of this has been shown both for a FRETpair based on nucleicacid base analogues [42] and for the FRETpairs Cy3 and Cy5 [43]. In both cases, thedecrease in FRETefficiency due to donor–acceptor separation along the DNA axis ismodulated by the orientation effect emanating from the turn of the DNA helix.Because most covalently bound fluorophores utilize flexible linkers to attach thefluorophore to the DNA, the orientation effect is not as apparent for all DNA-basedFRET systems. However, achieving a controlled orientation of the mediatingfluorophores that favors FRET can greatly enhance the efficiency of the wireconstruct. To illustrate, a change from randomly oriented donor and acceptors(k2¼ 2/3) to donor and acceptors with head-to-tail orientation (k2¼ 4) correspondsto a 35% increase in R0.An important practical consideration in the case of wires based on the homo-

FRET principle with a large number of mediator units connecting injector anddetector dyes is the probability for direct excitation of the mediating part of the wire.As wires grow in length, the number of mediator units needed to achieve end-to-endenergy transfer increases. Thus, the collective absorptivity of mediator part of thewire increases with increasing length. This means that if there is any probability fordirectly exciting the mediator dye at the injector excitation wavelength, the effect ofone-step transfer from the mediator to the detector dye will be more pronounced thelonger the wire is.

13.4.2Beyond Wires

The concept of linking functionalities on the nanometer scale using multistep FRETcan be expanded from the simple one-dimensional wire to include more complexstructures in two and (potentially) three dimensions. Multidimensional structuresprovide an opportunity to incorporate multiple functional units located within therange for multistep FRET. Here the positional precision of energy transfer becomes


even more important since there is a need to distinguish between different energytransfer targets.The development of DNA origami has allowed the creation of a wide variety of

DNA structures in two and three dimensions. DNA origami is based on the foldingof single-stranded genome of the M13mp18 bacteriophage into various shapes andpatterns using short staple strands. The technique was initially developed by PaulRothemund [44] and his work has later been followed by a many researchersimplementing the DNA origami concept in numerous ways [45–47]. The staplestrands used to assemble the origami structures also provide addressable handlesfor functionalizations [48]. This is utilized by Tinnefeld and coworkers to construct atwo-way photonic wire assembled in a DNA origami rectangle [49]. Here, an inputdye is placed at a central position in between two different output dyes, as shown inFigure 13.12a. Energy transfer from the donor to the acceptor is controlled using ajumper dye that functions as an intermediate station in the energy transfer cascadebetween the donor and both the acceptor dyes. Directional control is obtained byattaching the jumper dye to different staple strands at either positions between thedonor and the two acceptor dyes. In the design, the fluorophore ATTO 488 was usedas the donor dye, the fluorophores ATTO 647N and Alexa 750 were used as the twodifferent output dyes, and ATTO 565 was used as the jumper dye. The modifiedstaple strands are positioned diagonally across the origami rectangle with thefluorophores protruding from the same side of the flat structure. The structureis designed to give an average interfluorophore distance of roughly 9 nm. Theoreti-cally, this gives a low efficiency for direct donor–acceptor FRET in the absence of thejumper dye but a high efficiency for the two-step process in the presence of thejumper dye. The performance of the wire is studied using single-molecule fluores-cence spectroscopy. Without any of the two jumper dyes, the system is dominated byfluorescence from the donor dye with only a small fraction of the excitation energybeing transferred to the output dyes. When either of the jumper dyes is added, the

Figure 13.12 (a) Schematic image of the two-way photonic wire assembled on the DNA origamirectangle. (b) Energy transfer efficiency histograms for no, one, and two jumper dyes. Adaptedfrom Ref. [49].


excitation energy distribution is shifted toward the output dye on the side where thestrand with the jumper dye binds. This way, with the binding of a sequence-specificjumper strand, it is possible to control the directionality of the energy transfer.Single-molecule four-color FRET with alternating laser excitation was employed toprobe the energy transfer properties of individual constructs. When the greenjumper dye is positioned between the donor dye and the red output dye, the energytransfer efficiency, determined as the fraction of photon counts in the outputchannel divided by the sum of photon counts in the input and output channels,is 0.34 (Figure 13.12b). When the jumper dye is instead positioned between thedonor and the IR output, the energy transfer efficiency is centered at 0.27. Whenboth the jumper dyes are used simultaneously, the resulting energy transferefficiencies are 0.36 and 0.27 for energy transfer to the red and IR outputs,respectively. For all the constructs, the FRET efficiency distributions are quitebroad. Whether this heterogeneity is due to actual dynamics of the DNA origamistructure, or if it is caused by the photophysical properties of the fluorophores andthe fluorescence detection, is difficult to answer. In order to evaluate the readoutreliability of the system, an output ratio measure was defined as the photon countsfrom the red output channel divided by the sum of the red and IR outputs. When thejumper dye is located on the side of the red output, the output ratio is centered on 1,and when the jumper dye is located on the side of the side of the IR output, theoutput ratio is centered on 0, with little overlap between the two cases. Finally, whenboth jumper dyes are present, the output ratio is centered on 0.5. This shows that itis possible to differentiate between the different cases from the obtained signals.The key element in this work is, in principle, two combined photonic wires with acommon input, where selection is achieved by the addition of either jumper strand.In addition to this, the work also illustrates howmultiple chromophoric units can becombined in a complex matrix to perform a function. The binding fluorophore-labeled staple strands to the M13mp18 genome are separated from each other, butwhen the origami structure is folded, they are organized in a way that facilitates theenergy transfer from input to output.Another example of the expansion of the photonic wire concept to multi-

dimensional assemblies is presented by Albinsson and coworkers [50]. Here,fluorophores are attached to a hexagonal DNA assembly with arms extendingfrom the nodes of the hexagon. The DNA hexagon was designed by Tumpaneet al., with the aim to create a nanoscale DNA structure, with addressability on thesingle base pair level [51,52]. The hexagonal structure is built up of six three-waynodes, each consisting of three different 10-mer DNA oligos held together by a 1,3,5-tri-substituted benzene ring. The entire structure is spontaneously assembled whenthe different three-way nodes are mixed. In the assembled structure, the extendingarms as well as the sides of the hexagon, all have unique sequences, making itpossible to address them individually. The hexagonal structure is used to create aphotonic device with one input fluorophore and two, spatially and spectrallyseparate, output fluorophores (Figure 13.13a). Here, Pacific Blue functions asthe input fluorophore while fluorescein and Cy3 are the two output fluorophores.All three fluorophores are added to the assembly through covalent attachment to


10-mer oligonucleotides that bind to the single-stranded 10-mer extensions of thehexagon. Pacific Blue is a good FRETdonor to fluorescein but requires a mediatingchromophore to donate excitation energy to the other output dye, Cy3. Instead, amediating dye, in this case YO, is required for the Pacific Blue to Cy3 FRET to beefficient. This way it is possible to select which output in the multichromophoricsystem will be active; in the absence of YO, excitation energy is transferred tofluorescein (Figure 13.13b), while in the presence of YO, the excitation energy istransferred to Cy3 (Figure 13.13c).The results show that emission from the system, following excitation of Pacific

Blue, is dominated by fluorescein when no YO is present. As YO is added, Cy3emission becomes dominant (Figure 13.14a). The emission from Cy3 also increaseswith increasing YO concentration, with approximately 0.35 and 0.5 end-to-end FRETefficiency at 0.2 and 0.4 YO per base pair, respectively (Figure 13.14b). As in the case

Figure 13.13 (a) Multichromophoric systemwith one input and two, spectrally and spatiallyseparate, outputs. (b) In absence of themediator dye, the system favors energy transfer

to fluorescein. (c) Addition of YO leads toenergy transfer from Pacific Blue to Cy3.Adapted from Ref. [50].

Figure 13.14 (a) Difference in Cy3 and fluorescein emissions for different YO to base pair ratio.(b) Experimental and simulated estimates of Pacific Blue to Cy3 FRET as function of YO to basepair ratio. Adapted from Ref. [50].


of the photonic wire comprising the fluorophores Pacific Blue, YO, and Cy3 (seeabove), the excitation energy that is transferred between Pacific Blue and Cy3 ismediated by YO in a diffusive manner. The difference is that the fractal dimen-sionality of the hexagonal assembly contains more stray paths where excitationenergy can be lost, than the one-dimensional wire, where the excitation energy cantravel only along the path connecting the initial donor and the terminal acceptor.Therefore, there needs to be a strong connection, that is, very efficient diffusiveenergy migration, for end-to-end energy transfer to be efficient. Thus, comparedwith the wire, the hexagon requires a high loading of mediators for the Pacific Blueto Cy3 energy transfer to be efficient. The contrast between the state dominated byfluorescein and the state dominated by Cy3 emission is diminished to some extentbecause of two factors: direct energy transfer between Pacific Blue and Cy3, andenergy transfer between fluorescein and Cy3. In addition to the fluorescencemeasurements, the multichromophoric system was also investigated using theMarkov chainmodel described above. The simulations show that the energy transferbehavior of the system is distinctly different at different YO concentrations. Thearrival rate of excitation energy at Cy3 shows a strong dependence on the mediatorloading with an almost immediate steep increase in excitation energy at Cy3 at 0.4YO per base pair. In contrast, at 0.1 loading ratio, there is a 1 ns time lag beforeexcitation energy starts to arrive at Cy3.The above examples show how the concept of the photonic wire can be incorpo-

rated intomore complex assemblies to introduce new functionalities. Both examplesdeal with control of directionality of the flow of excitation energy through thepositioning of fluorophores in the DNA assembly. However, the added complexitycan also be used to include functions that relate to the output and usage of thetransferred excitation energy, as well as the design of systems for maximizing theinflux of excitation energy to specific positions. Such applications will be discussedin the following sections of the chapter.

13.4.3Light Harvesting

In addition to transferring energy from one specific point in space to another (as inthe case with the photonic wire), there are also examples were multichromophoricFRET has been employed to convey excitation energy from a large number offluorophores to few or single acceptors. The inspiration here comes of course fromthe light-harvesting complexes in photosynthetic organisms, described in theprevious section. The approach that is used in the artificial light-harvesting devicesresembles in a way, a funnel for excitation energy, as excitation energy at any donorfluorophore is directed toward the acceptor or acceptors in a way schematicallydepicted in Figure 13.15. This type of funnel, or antenna assembly, can be created innumerous ways and can function in either one [53], two [54], or three dimensions[55]. The molecular foundation of the antenna complexes has varied greatlyand examples include cyclodextrins [56], dendritic [57,58] or ring-shaped [58–63]arrangements of porphyrins, and pyrene–perrylene assemblies [64,65]. The


assemblies rely either on scaffolding of chromophores around some kind ofmolecular framework or on the ability of the chromophores themselves to arrangeinto an antenna complex.Especially the ring-shaped porphyrin assemblies closely mimic the arrangement

found in the bacterial light-harvesting complexes LH-I and LH-II. All these examplescan be discussed in length. However, the focus of this section is primarily onantenna assemblies relying on self-assembly of biomolecules, for example, DNA orproteins, into scaffolds for multichromophoric arrangements similar to those usedto construct the photonic wires previously discussed. We have chosen to make thisrestriction in order to be able to describe each example in detail within the limitationof a single chapter. In addition, we want to emphasize the connection betweennanoscale photonic devices, such as the photonic wire, DNA nanotechnology, andlight-harvesting assemblies. The idea is that these light-harvesting applicationscould function in conjunction with other nanoscale devices, for example, photonicwires, present within the DNA assemblage.As in the case of the photonic wire, it is possible to find examples that have the

properties of a light-harvesting device, although they might not be branded as such.Armitage and coworkers have created a set of two- and three-dimensional DNA-based nanostructures that function as hosts for the intercalating fluorophore oxazoleyellow (YO-PRO-1) or its bisintercalating dimer YOYO-1 [54,55]. The intercalatedYO-PRO or YOYO dyes act as FRETdonors to an acceptor fluorophore that is eithercovalently bound to the DNA structure or intercalated between the base pairs. Theassemblies effectively work as nanoscale fluorescent markers, organizing multiplefluorophores in a defined region of subdiffraction size. Although the DNA-basedassemblies are primarily designed as brightly fluorescent tags, they are interesting todiscuss from a more generalized light-harvesting perspective. In 2007, Armitageand coworkers published work on the creation of a two-dimensional DNA-basedfluorescent nanotag [54]. The DNA construct that functions as scaffold for theintercalating dyes is a three-way junction (3WJ) with “arms” that are 10 base pairslong. In total, three 20 base pair long, single-stranded DNA molecules are used toconstruct the 3WJ. The 3WJ assembly was constructed in two versions, one with an

Figure 13.15 Schematic illustration of the energetic principles behind light-harvesting devices. Bymeans of hetero- and homo-FRET, excitation energy is transferred from multiple copies of a donordye to a single copy of an acceptor dye.


intercalating acceptor (Figure 13.16a) and one with covalently attached acceptormolecules (Figure 13.16b). Another cyanine dye, TO-PRO-3, was used as theintercalating acceptor. As with the case of the intercalator-based photonic wire,there is no control of the positioning of the fluorophores when intercalating donorand acceptor dyes are used. Instead, the ratio between added donor and acceptor isused in order to achieve high efficacy for both the light-gathering and energytransfer processes. A 37% quenching of YO-PRO-1 is reported for donor/acceptorratio of 14 : 1. An increase in the acceptor concentration to ratios 13 : 2 and 12 : 3results in 53 and 67% quenching, respectively. However, there is no increase in thesensitized emission of the acceptor with increasing donor quenching. A probableexplanation for this is self-quenching of TO-PRO-3 due to the high loading;something that is supported by decreasing fluorescence intensity from directlyexcited TO-PRO-3 with increasing loading.The alternative strategy, with covalently attached acceptors, involves YOYO-1 as

donor and Cy3 as acceptor. Cy3 is attached to the terminus of the 3WJ assembly. Anadvantage with the covalent attachment is that the acceptor dyes do not compete withthe donor dyes for the available binding sites in the DNA structure. The alternativeassembly with covalently attached acceptors showed 95% quenching of the YOYO-1donor for a structure with two Cy3 acceptors. Using this as an example, it wouldappear as if covalent attachment of the acceptor is preferable over noncovalentbinding. Certainly, if the assembly should function as a light-harvesting device witha focused output, there should only be one acceptor, placed in a well-definedposition. With that in mind, the covalent option is more suitable. It is not asimportant to precisely control the positioning of the donors. As long as the bindingsites are within energy transfer distance to the acceptor, random binding of thedonors will not hamper the light-harvesting functions.A further development of the DNA-based fluorescent nanotag is made possible

using the DNA tetrahedron [55], introduced by Turberfield and coworkers [66]. Thesame donor–acceptor pair as for the three-way junction, YOYO-1 and Cy3, is usedfor the tetrahedron (TH). Each side of the tetrahedron is 17 base pairs in length,

Figure 13.16 Light-harvesting three-way junctions with intercalated (a) and covalently attached(b) acceptor dyes. Adapted from Ref. [54].


connected with two base pair hinges (Figure 13.17). This means that there are a totalof 102 binding sites for YOYO-1 in the structure, compared with 30 for the 3WJ,without greatly increasing the size of the assembly. The increased compactness ofthe TH structure, and the concentration of donors that follows, is a distinctadvantage with using three-dimensional assemblies over one- or two-dimensionalones. Using a YOYO-1 loading density of approximately 1 : 4 YOYO-1: base pair,Armitage and coworkers obtained donor quenching efficiencies of 52, 74, 86, and95% for structures with one, two, three, and four Cy3 acceptors, respectively [55].Apoint that is not being discussed in either of the two papers is the effect of homo-

FRETon the light-harvesting capacities of the assembled structure. Would a similarconstruct, built using donor fluorophores without homo-FRET capabilities, showsubstantially lower sensitized acceptor emission? Homo-FRET should, in principal,play an important role to direct excitation energy from donor dyes distal from theterminal acceptor toward the acceptor through diffusive energy migration.The question of the importance of homo-FRET is, in part, addressed by Francis

and coworkers in the creation of a protein-based, multichromophoric, antennasystem [67]. Recombinant tobacco mosaic virus coat protein (TMVP) monomers areorganized into either rods or disk shapes through self-assembly. The proteinmonomers are S123C mutants, that is, a serine residue is substituted for cysteine.This provides a handle for attachment of thiol-reactive fluorophores. The overalldesign involves three different fluorophores, Oregon Green 488 (A), tetramethylr-hodamine (B), and AlexaFluor 594 (C). The fluorophores can be arranged into aFRET cascade, where A is the initial donor, B the intermediate donor, and C the

Figure 13.17 Light-gathering tetrahedron constructed by Armitage and coworkers using YOYO-1as donor and Cy3 as acceptor. Adapted from Ref. [55].


acceptor. The extent of overlap between the emission band of the donor and itsabsorption band allows for substantial donor–donor FRET (homo-FRET). Theassembly of the TMVP monomers into the final structure is directed by pH; alow pH favors the formation of rods while a high pH results in the formation ofdisks. Experiments with assemblies containing only fluorophores A and C showedthat energy transfer from donor fluorophores was the main contributor to acceptoremission at high donor: acceptor ratios (Figure 13.18a). A comparison between a rodassembly carrying 33 donors for each acceptor with a similar assembly, but withequal amounts of donors and acceptors, showed that at least 20 donor fluorophorescontributed to the acceptor emission. This is more than the maximum eight donorsthat can be positioned directly adjacent to the acceptor. Although it is possible thatthe contribution from nonadjacent donors comes from direct FRET, the multistep,homo-FRET pathway presumably has an important contribution. To evaluate thelight-harvesting capabilities, Francis and coworkers measure the antenna effect ofthe system. The antenna effect AE was introduced by Fr�echet and coworkers todescribe the functioning of dendritic systems [68] and is defined here in Equa-tion 13.7,

AE ¼ IA d exð Þ=IA a exð Þ; ð13:7Þwhere IA is the emission intensity of the acceptor, d_ex signifies donor excitation,and a_ex direct acceptor excitation, both excited at the respective absorptionmaxima. In order to function as an estimate of the light-harvesting capabilitiesof a system, the compared emission intensities must be corrected for differences inillumination. In the mentioned example, the antenna effect increases with increas-ing donor: acceptor ratios from AE¼ 0.3 at 1 : 1 ratio to a fourfold amplification(AE¼ 4) at 16 : 1 ratio. A further increase in the donor: acceptor ratio to 33 : 1 and101 : 1 results in an eightfold and 11-fold increase, respectively. Although a highantenna effect is observed, the overall efficiency of the system is fairly low, rangingfrom 34% to a maximum of 47%. This is explained by the poor spectral overlapbetween the donor emission and the acceptor absorption. Including the intermedi-ate donor greatly increases the overall efficiency. By first constructing disk structuresthat contain either fluorophores B and C or only fluorophore A, and letting theseself-assemble to a rod structure, it is possible to position the intermediate donor in-between fluorophores A and C, thereby obtaining a fluorophore organization

Figure 13.18 Two- (a) and three-color (b) light-harvesting devices based on modified tobaccomosaic virus coat protein. Adapted from Ref. [67].


promoting end-to-end FRET (Figure 13.18b). At fluorophore ratio 8 : 4 : 1, the overallefficiency is 90% with an antenna effect of 4.6.Another way to spatially organize donor and acceptor fluorophores in order to

promote light gathering is presented by Liu and coworkers, who have constructed anartificial light-harvesting antenna from a seven-helix DNA bundle [69]. Donorfluorophores are positioned in six helixes surrounding a central helix that housesthe terminal acceptor fluorophore, schematically illustrated in Figure 13.19a. Threedifferent fluorophores are used in the assembled structure: ethynylpyrene (Py), Cy3,and AlexaFluor 647 (AF). Pyrene functions as a primary donor to the intermediatefluorophore Cy3, which in turn functions as a donor for the acceptor fluorophoreAlxaFluor 647. The circular arrangement of donors around the acceptor creates aneffective funnel for the excitation energy directing into to the acceptor. This circularorganization resembles the organization of the bacterial light-harvesting complexLH-I around the reaction center. The self-assembled seven-helix bundle (7HB) wasinitially designed by Seeman and coworkers [70].Four different assemblages with different chromophoric conformations were

constructed. The chromophore assemblies, or triads, have the following dye ratios(donor: intermediate donor: acceptor): 6 : 6 : 1, 6 : 3 : 1, 3 : 6 : 1, and 1 : 1 : 1 for thetriads T1, T2, T3, and T4, respectively (Figure 13.19b). The chromophores (with theexception of the central AF acceptor) are held rigidly close to the DNA helix with a

Figure 13.19 (a) Three-chromophoric light-harvesting device based on a seven-helix DNAbundle. (b) Illustration of the chromophoric arrangement in the 6 : 6 : 1, 6 : 3 : 1, 3 : 6 : 1, and 1 : 1 : 1triads. Adapted from Ref. [69].


relatively well-defined orientation. Because of this, the assumption of randomorientation (k2¼ 2/3) does not hold. It also means that donor–acceptor distancesare relatively well defined (especially for the Py–Cy3 pair). For the T1 assembly(6 : 6 : 1), Py–Cy3 distances range between 2.1 and 2.7 nm. The Cy3–FA distancevaries slightly more than the Py–Cy3 with a minimum distance at 1.8 nm to amaximum distance at 4.5 nm.When the assemblies are excited at 380nm, there is a quenching of the donor

fluorophore Py for all the triads, T1–T4, comparedwith an assemblywithout any of thetwo acceptors Cy3 and AF. The T1 (6 : 6 : 1) and T3 (3 : 6 : 1) show a 90% quenching ofPy, while T2 (6 : 3 : 1) and T4 (1 : 1 : 1) show 30 and 70% quenching, respectively.Comparing the results of the T1 and T2 triads, it is clear that a high ratio ofintermediate donors is required to achieve an effective donor quenching. However,increasing the amount of intermediate donors beyond a 1 : 1 ratio does not lead tofurther quenching of the primary donor (T1 and T3). Finally, the difference in donorquenching between T1 andT4 shows that there is an effect ofmultiple energy transferpaths directing the excitation energy away from the Py donor.As was the case with the photonic wire, it is not sufficient to just study the

quenching of the primary donor to fully characterize the performance of the antennasystem. Using the antenna effect measure introduced by Francis and coworkers [67],the efficiency of energy transfer to the terminal acceptor AF is estimated. Thereported antenna effects for the T1, T2, T3, and T4 triads are 0.85, 0.43, 0.47, and0.16, respectively. It is interesting to note that the antenna effects of the T2 and T3triads are approximately half that of T1. T2 has the same number of primary donorsas T1 but only half the number of intermediate donors. Thus, T2 absorbs the sameamount of light as T1, but the relaying is more inefficient. This is also reflected inthe lack of quenching of the primary donor. T3 relays the excitation energy with thesame efficiency as T1, but absorbs only half the amount of light that T1 does.

13.4.4Functional Control

A lot of work has been put into creating molecules that can be switched between twodifferent states in a controllable way. Being it through light, changes in pH, or someother stimuli, a switching mechanism provides a means to regulate a function, forexample, by turning it on and off. In the case of the photonic wire, one mightconsider switching the wire between a conductive and a nonconductive state.Ideas on how to regulate wire performance were proposed by Lindsey and

coworkers in 1996 [26], shortly after the construction of the original porphyrinmolecular photonic wire in 1994 (see above) [25]. The main constituents of the wireare the same as those mentioned above: a boron-dipyrromethene (BODIPY) input, aZn porphyrin (ZnP) as linker, and a free-base porphyrin as output. In addition, thereis a possibility to control the wire performance by switching the redox state of ametalloporphyrin to generate a stable radical cation. To serve as a switch for thephotonic wire, four criteria were put up for the metalloporphyrin. The first criterionstates that the absorption energy of the neutral metalloporphyrin should be higher


than that of the free-base porphyrin. In the “conducting” form of the wire thereshould be no tendency for the excitation energy to take the “nonconducting”pathway. In contrast, the second criterion states that the metalloporphyrin shouldhave low-lying energy levels in its oxidized state. This state should also be non-fluorescent. The requirement for nonfluorescence is important when consideringthe metalloporphyrin as a switch turning the transmission of the wire on and off. Itis also possible to view the switch as a tool to change the directionality of the “wire”directing the excitation energy to a secondary output. However, the way that Wagneret al. [26] have designed their wire stipulates the use of themetalloporphyrin as an onand off switch. The third criterion states that the redox properties of the metal-loporphyrin should be reversible, providing the basis for its function as a switch.Finally, the fourth criterion states that the metalloporphyrin that is being switchedshould have the highest energy highest occupied molecular orbital (HOMO) of allthe units in the assembled supramolecular structure. This ensures that the oxi-dization occurs at the correct site. In the work by Wagner et al., a magnesiumporphyrin, fulfilling all the above stated criteria, is selected to function as the redoxswitch, regulating the functionality of the photonic wire. Figure 13.20 shows theenergy level diagram of the components included in the array.Two important considerations when it comes to functional control through

switching between different states are switching efficiency and switching rate. Ifwe apply a device orientation in the analysis of switching mechanisms, we are facedwith certain restrictions. While a molecular switch can, in principle, have a turnoverrate of several hours, this is not compatible with the way we generally conceive theword device. The word device has certain connotations, one of them being that thespeed of operation is in the order of seconds, or even fractions of seconds. Thiseffectively limits the number of applications that can be considered as switchingdevices in some kind of technological sense, more so when the switch in questionshould be of molecular nature, regulating the functionality of a multichromophoricassembly such as a photonic wire. There is no switching rate reported in the

Figure 13.20 Energy transfer cascade going from a boron–dipyrromethene unit, through a Znporphyrin, to the free-base porphyrin output. The redox state of a Mg porphyrin regulates thefunctionality, switching the wire between an ON and an OFF states. Adapted from Ref. [26].


work by Lindsey and coworkers. However, the switching is highly efficientwith 99.9% oxidation of the Mg porphyrin, with only about 4% oxidation of theZn porphyrin. The oxidation is achieved by means of electrochemistry, but, asLindsey and coworkers hint in the paper, it might also be possible to achieve itthrough photochemistry. We will return to photochemical switching later on inthis section.As stated above, one method for changing, for instance, the binding and

unbinding of a functional unit to a DNA structure is the regulation of pH. Inthis example by Tumpane et al., the binding of a triplex strand to a nanoscale DNAstructure is controlled by changes in pH [51]. The association of the triplex strand isfollowed by the switching of an energy transfer process involving fluorophores onthe triplex forming single strand and the DNA nanoconstruct. Changes in pH can bea very effective way to control the association of auxiliary molecules to DNA.However, one issue that is common for all devices relying on chemical switchingto control functions is the question of dilution. To avoid this, other means offunctional regulations can be considered; we have already mentioned switching byelectrochemical means. Another way to obtain this functional regulation is to usemolecules that can isomerize when exposed to light of certain wavelengths. As weare already discussing photonic devices, stimulating our system using light comesnaturally. We will limit this discussion to instances where photoswitching is ofrelevance to DNA-based photonic devices.Regulation of a photonic device by means of photoswitching can be divided into

two principal cases. A molecule can be isomerized between a binding and anonbinding state. This way, the association of a mediating fluorophore in, forexample, a photonic wire can be controlled, regulating the function of the wire. Inthe other case, functional control is added by the introduction of photochromicmolecules as part of the multichromophoric system. The photochromic moleculeswitches between a bright and a dark state and can, in this way, be used to turn amulti-FRET cascade on and off.Komiyama and coworkers have investigated the possibility to use light to control

the melting of a DNA duplex using a DNA strand modified with an azobenzenemoiety [71,72]. Upon irradiation by UV light (300 nm< l< 400 nm), the azoben-zene is isomerized from trans to cis form. The trans! cis isomerization induces alowering of the melting temperature for short DNA duplexes. This process isreversed by the exposure to light at wavelengths above 400 nm. The extent of thedecrease in melting temperature varies with DNA sequence [72], but for the meltingtemperature to sink below room temperature, the modified sequence needs to bepolyadenine [71]. Another option that has been explored is to incorporate theazobenzene in a separate DNA strand that is associated with a target duplex throughtriplex binding [73]. The advantage of using the triplex strategy is that it enables theduplex target to be part of a larger structure. This way, it would be possible toincorporate this triplex switch in the larger DNA assemblies previously discussed. Arelated example, which does not require covalent attachment to a single-strandedDNA to display light-controlled association and dissociation to DNA, comes fromthe spiropyran family [74].


13.4.5Quantum Dots in Multistep FRET

So far, we have mostly considered photonic devices built using conventional organicfluorophores. However, new and interesting functions can be obtained by expandingthe scope to include novel dyes. Quantum dots (QDs) have emerged as aninteresting alternative to conventional organic dyes, lacking some of the issuesassociated with organic dyes, such as susceptibility to photobleaching or photo-degradation, or broad absorption and emission bands [75]. Quantum dots have alsoseen a widespread use in biological and imaging applications [76–78]. Medintz andcoworkers have constructed DNA-based photonic wires where quantum dots areassociated with the duplex DNA and act as the initial donor in a multichromophoricFRET cascade [79,80]. The photonic wire is assembled around a 40-mer scaffoldstrand. The different acceptor dyes are added to the assembly attached to short 10-mer strands that hybridize with the scaffold. The short strands bind to the scaffold sothat the separation between the different acceptor dyes is 10 base pairs, ensuringthat they protrude from the same side of the DNA helix. The association of thequantum dot to the DNA scaffold is enabled by the metal affinity coordination of ahexahistidine peptide that is attached to the end of one of the strands. The histidinescoordinate to the Zn-rich surface of the CdSe–ZnS core–shell QDs that are used.The assembly of the QD-associated photonic wire is outlined in Figure 13.21. Withan emission maximum around 530nm, the quantum dot acts as a FRET donor toCy3 that, in turn, forms a FRET pair with Cy5. The wire can be further extended toinclude the dyes Cy5.5 and Cy7. Because of the size of the quantum dots, it ispossible to coordinatemore than one (His)6 unit to one QD, and thus, it is possible tocreate a number of photonic wires sharing a common initial donor. In a sense, this isthe opposite process to that of the light-harvesting devices. Instead of concentratingthe excitation energy by going from multiple, spatially separate, donors to a single,centrally located, acceptor, we have an expansion of the excitation energy withexcitation at one single donor reaching multiple targets. Using only the QD–Cy3donor–acceptor pair, Medintz and coworkers investigated the dependence of theFRETefficiency on the number of coordinating acceptors. This showed an increasedquenching of donor emission, with a simultaneous enhancement of the acceptorfluorescence. A ratio of four wires per quantum dot was selected for use in thedifferent subsequent wire assemblies. As the organic dyes are tethered to 10 baselong oligomers, there are four distinct fluorophore positions (denoted 1–4) along the40-mer scaffold.In an initial conformation, Cy3 was located at position 1, while Cy5 was placed

both at positions 2 and 3. Addition of Cy3 to the assembly leads to a pronouncedquenching of the QD emission. However, only about 25% of the initial excitationenergy at the QD is emitted by Cy3. When Cy5 is added to position 2, there is a 60%quenching of Cy3, with little or no change in the QD emission intensity. Addition ofCy5 to position 3 leads to a further decrease in the Cy3 fluorescence by about 20%. Asecond confirmation is contained the dyes Cy3, Cy5, and Cy5.5 in positions 1, 2, and3, respectively. In addition to this, the intercalating dye BOBO-3 was used. BOBO is a


dimeric cyanine dye, similar to YOYO, which inserts between the base pairs of DNA.BOBO has, just as YOYO, the ability to mediate excitation energy transfer throughhomo-FRET. In this system, BOBO would primarily acts as acceptor in pair with thequantum dot, as both donor and acceptor with Cy3, and as donor together with Cy5and Cy5.5. However, although BOBO-3 effectively quenches both the quantum dotand Cy3, there is no energy transfer to the downstream acceptors Cy5 and Cy5.5. It islikely that the BOBO-3 dyes create energy sinks in the system, preventing efficientend-to-end transfer. Finally, a conformation containing Cy3, Cy5, Cy5.5, and Cy7 inpositions 1, 2, 3, and 4, respectively, was tested. In this assembly, the quantum dot isquenched to 80–90%. Quenching is also efficient for Cy3 and Cy5. Without, andwith, Cy5.5 as terminal acceptor, approximately 1% of the total available excitationenergy is emitted from Cy5.5. Addition of the Cy7 strand leads to a decrease in theCy5.5 emission to 0.3%, without any detectable increase in the Cy7 fluorescence.

Figure 13.21 Assembly of quantum dot-associated photonic wire. Adapted from Ref. [80].


13.4.6Potential Outputs and Uses for Channeled Excitation Energy

The excitation energy that is channeled in the wire system has to be directed to somefunctional group to be made of use. In general, the output can be a molecule capableof undergoing reactions in its excited state. This way, it is possible to controlchemical reactions using the different techniques described in this chapter. Onestrategy, similar to what occurs in the photosynthetic reaction center, is to convertthe excitation energy to charge separation through a photoinduced electron transferreaction. As is the case in the bacterial reaction center, the electron transfer reactioncan be coupled to the generation of a chemical potential across a membrane,providing a driving force for a multitude of chemical reactions.In energy transfer cascades, porphyrins can not only act as excitation energy donor

and acceptor but also as a functional group capable of performing photoinducedelectron transfer. Albinsson and coworkers have constructed a supramolecularassembly where a Zn porphyrin is coupled to a DNA strand by a linking moiety[81,82]. The porphyrin has two principal functions in the supramolecular assembly:it acts as an anchoring group associating the DNA porphyrin assembly to a lipidbilayer, and functions as an electron donor participating in electron transfer withquinine molecules buried in the hydrophobic part of the bilayer (Figure 13.22a). Intheir first design, the ZnP was attached to five base pairs from the 50 terminus in a14-mer DNA strand. In either end of the double-stranded DNA, there is afluoresceinmoiety attached. Fluorescein functions as a FRETdonor to the porphyrinunit. Since the porphyrin unit is not centered between the two fluorescein moieties,two different energy transfer efficiencies could be detected: 85 and 65% for the shortand long distance, respectively. The assembly is bound to 100 nm diameter lipo-somes, using the ZnP as an anchor. Since energy transfer occurs between fluores-cein, which protrudes into the solution, and the porphyrin, which is buried in themembrane, the assembly introduces a phase transition, where excitation energy istransferred from an aqueous phase to a lipid phase. The sequential energy transferand electron transfer experiments were performed using 2,6-di-tert-butyl-p-benzo-quinone (BQ) buried in the lipid bilayer quenching the ZnP emission throughelectron transfer. The quenching process was initially studied using direct excitationof the DNA–ZnP without any attached fluorescein. The BQs were shown toeffectively quench the ZnP fluorescence and a diffusion controlled bimolecularquenching constant of 1.8� 109 M�1 s�1 was obtained. The experiments usingexcitation of fluorescein instead of ZnP approximately showed the quenchingresults as for the porphyrin-only experiments. Thus, the quenching of porphyrinfluorescence is independent of whether the excitation energy comes from FRET orfrom direct excitation. Further work has explored hexagonal DNA structures wheremultiple porphyrin anchors are used to anchor the assemblies in the lipid mem-brane (Figure 13.22b) [83] and dimeric zinc porphyrin pockets as binding sites forbidentate ligands [84].The above example fits very well with the previously described work relying on

DNA as a scaffold for fluorophores in the creation of photonic devices, including


photonic wires and light-harvesting devices. However, a lot of work has been put intothe creation of reaction center-like structures, where excitation energy is gatheredand utilized to drive electron transfer reactions, using conventional organic mol-ecules. We will mention some of these works since they have great conceptualimportance in the design of nanoscale photonic devices.Devens Gust, Thomas A. Moore, and Ana L. Moore have since long worked with

the aim of creating molecular assemblies mimicking energy and electron transfer inphotosynthesis [85–88]. In their work, they construct artificial reaction centers wherethe key components are one porphyrin (P) and one fullerene (C60). In the simplestcase, the artificial reaction center consists only of the porphyrin and the fullerene,functioning as electron donor and acceptor, respectively (Figure 13.23a). The two

Figure 13.22 (a) Supramolecular assembly ofDNA, Zn porphyrin, and fluorescein anchoredto a lipid bilayer. Energy transfer fromfluorescein to the porphyrin enables excited-state electron transfer to the quinones buried in

the hydrophobic part of the bilayer. Adaptedfrom Ref. [81]. (b) DNA nanostructure withthree-porphyrin moieties linking the assemblyto a lipid bilayer. Adapted from Ref. [83].


units are covalently linked to provide electronic coupling to facilitate photoinducedelectron transfer between the two moieties. In this simple case, excitation of theporphyrin to its first excited singlet state leads to photoinduced electron transfergenerating P�þ–C60

�� with unit quantum yield. The rate constant for this process is3.3� 1011 s�1. A limitation to this simple design is the fast decay back to the groundstate through charge recombination with a lifetime of 478 ps. As in the naturalreaction center, a more long-lived charge-separated state can be achieved byincreasing the separation between the generated charges. A larger distance leadsto a weaker electronic coupling and a slower charge recombination. A way to obtainincreased separation is to add a carotenoid (C) secondary electron donor to theprevious dyad molecule, creating a C–P–C60 triad (Figure 13.23b) [89]. Hereexcitation of the porphyrin again results in photoinduced electron transfer yieldingC–P�þ–C60

��. In this state, there is a competition between the previously describedcharge recombination and electron transfer from the carotenoid to the porphyrinwith the product C�þ–P–C60

��. The generation of C�þ–P–C60�� proceeds with a

quantum yield of 0.88. From this state, charge recombination is much slower andthe lifetime of the charge-separated state is extended by a factor of nearly 1000,relative to the dyad. A design that closely relates the artificial light-harvesting devicespreviously discussed is obtained by replacing the carotenoid moiety by an artificialantenna, comprising four zinc tetraarylporphyrins [88]. The energy transfer designresembles that of the photonic wires created by Lindsey and coworkers (see above)[25,26]. Another step in the development of biomimetic reaction center-like devices

OC

O

N

N

N

N

HH N

CH3

N

H

C

O N

N N

N

HH N

CH3

(a)

(b)

Figure 13.23 (a) Charge separating dyad comprised of a free-base porphyrin electron donor anda fullerene electron acceptor. (b) The dyad extended to a triad with an electron-donatingcarotenoid. Adapted from Ref. [87].


has been the development of a light-driven transmembrane proton pump [90]. Thisinvolves a replacement of the fullerene from the triad with a quinone to create acarotenoid–porphyrin–quinone (C–P–Q) triad. The pump is assembled by vectorialinsertion of the triad into vesicle in a way that places the quinone near the externalsurface. The pump functions as a redox loop, with the triad as the key component.Upon irradiation, the triad forms the charge-separated state C�þ–P–Q�� with closeto unit efficiency. The radical anion near the membrane exterior can then reduceanother quinone in the solution producing a semiquinone. The neutral semi-quinone can readily diffuse across the membrane to the interior surface where itencounters the carotenoid radical cation. The result of this encounter is that thesemiquinone is reduced back to a quinone, and in this process releases a proton intothe interior of the liposome.Another example where the bacterial reaction center is used as an inspiration to

create devices for charge separation comes from W€urthner and coworkers [64]. Inthis work, a multichromophoric structure is created comprising 16 pyrene chro-mophores tethered to a square-shaped perylene bisimide scaffold (Figure 13.24).Steady-state fluorescence measurements show that the pyrene moiety is stronglyquenched. This could indicate either energy transfer from pyrene to perylene or anelectron transfer process from pyrene to the perylene electron acceptor unit. Asidefrom this, there is also quenching of the perylene bisimide emission due to electrontransfer. Time-resolved fluorescence measurements show a 178 ps rise time in theemission from the perylene unit, which is in fair agreement with themain quenchedlifetime of pyrene at 231 ps. This suggests that energy transfer occurs from pyrene toperylene. Comparison with unquenched pyrene in nonpolar solvent gives a rateconstant for the energy transfer process of 5.0� 109 s�1, which corresponds to a

donor–acceptor distance of 8–11A�. The quenching of perylene, seen in the steady-

state measurements, was observed as a reduction in fluorescence lifetime, from5.5 ns for the perylene square without pyrene moieties, to 930 ps for the completepyrene–perylene assembly. The nature of the quenching of the perylene unit wasfurther investigated with femtosecond transient absorption. The transient absorp-tion measurements show the following processes occurring: instantaneous excited-state population of pyrene followed by a depopulation of the perylene bisimideground state with a time constant of 120 ps. This process is indicative of energytransfer between pyrene and perylene. In addition, absorption bands form at 470and 780 nm. These two peaks correspond to pyrene radical cation and perylenebisimide radical anion formation, respectively, showing that rapid electron transferfrom pyrene to perylene bisimide occurs alongside energy transfer. Finally, theperylene bisimide radical cation band decays due to the back electron transfer.Compared with the individual ligand consisting of four pyrene units coupled to theperylene bisimide moiety, the electron transfer process is accelerated in the square-shaped assembly. In addition to this, the efficiency is also increased from 70 to>94%. This is attributed to lowering of the LUMO level of the perylene bisimide,rigidification caused by the steric crowding in the square and an increase in the localchromophore concentration in the square compared with the individual ligand.W€urthner and coworkers suggest that the multichromophoric square, consisting of


20 chromophoric units packed in a very small volume, might be considered as aninterfacial structure approaching the solid state.

13.5Summary

The interest inmultichromophoric FRET has grown together with both expansion ofthe field of self-assembled nanotechnology and the development of even moresophisticated fluorescence-based analysis techniques. In this chapter, we havediscussed the fundamental principles behind multichromophoric FRET and putit in relation to its use as a technique to gain information on molecular structurebeyond the linear distance provided by the two-color FRET. We have also briefly

Figure 13.24 Light-harvesting square seen from the top (a) and from the side (b). Pyrenes aredepicted in blue and perylene bisimides in red. Adapted from Ref. [64].

13.5 Summary j647

described the light-harvesting system in bacterial photosynthesis and how it directsexcitation energy to the reaction center through energy transfer in multiple steps.Finally, we have discussed the growing field of self-assembled DNA nanotechnologyand howmultistep FRET functions in this field as an information carrier in photonicwires with a reach above 20 nm and with lower nanometer resolution. Here we havealso touched upon the further development of the wire concept to includemore thanone dimension. Recent work has shown great potential for the wire concept to beincluded in more complex DNA-based systems as well as potential for the conven-tional organic fluorophores to be accompanied by quantum dots in the long-rangeenergy transfer system. Channeling the excitation energy to predetermined sites ona DNA nano network is not the final goal. Once transferred, the excitation energyshould do some work either through inducing an excited-state reaction or bymimicking the charge separation reaction of the natural photosynthesis. A fewattempts along these lines were also presented at the end of this chapter.

13.6Note Added in Proof

Since writing this chapter, one important piece of work has been carried out, andpublished [91], which add important knowledge for our understanding of themultistep FRET systems discussed in this chapter.The concept of homo-FRET was used in a self-assembled DNA-based artificial

antenna complex. By using intercalated YO-chromophores excitation energy wascollected and transported to a final acceptor (the “reaction center”) consisting of asingle free-base porphyrin covalently attached to one of the DNA-strands. Theconcept was demonstrated both for linear and branched hexagonal DNA constructsand antenna effects reaching 12 and 18 were realized for the respective constructs.This study points to a simple way of constructing self-assembled nano-scale lightharvesting antennas that mimic the natural photosynthetic antennas.

References

1 Perrin, F. (1932) Polarization offluorescence. Annales de Physique, 17, 283.

2 Pheofilov, P.P. and Sveshnikov, B.J. (1940)On the concentrational depolarization ofthe fluorescence dye-stuff solutions.Journal of Physics-USSR, 3, 493–505.

3 Weber, G. (1954) Dependence of thepolarization of the fluorescence on theconcentration. Transactions of the FaradaySociety, 50, 552–560.

4 F€orster, T. (1948) Zwischenmolekulareenergiewanderung und fluoreszenz.Annalen Der Physik, 2, 55–75.

5 Stryer, L. and Haugland, R.P. (1967)Energy transfer – a spectroscopic ruler.Proceedings of the National Academy ofSciences of the United States of America, 58,719–724.

6 Watrob, H.M., Pan, C.P., and Barkley, M.D.(2003) Two-step fret as a structural tool.Journal of the American Chemical Society,125, 7336–7343.

7 Tinnefeld, P. and Sauer, M. (2005)Branching out of single-moleculefluorescence spectroscopy: challenges forchemistry and influence on biology.


Angewandte Chemie, International Edition,44, 2642–2671.

8 Kapanidis, A.N., Lee, N.K., Laurence, T.A.,Doose, S., Margeat, E., and Weiss, S. (2004)Fluorescence-aided molecule sorting:analysis of structure and interactions byalternating-laser excitation of singlemolecules. Proceedings of the NationalAcademy of Sciences of the United States ofAmerica, 101, 8936–8941.

9 Lee, N.K., Kapanidis, A.N., Koh, H.R.,Korlann, Y., Ho, S.O., Kim, Y., Gassman,N., Kim, S.K., and Weiss, S. (2007) Three-color alternating-laser excitation of singlemolecules: monitoring multipleinteractions and distances. BiophysicalJournal, 92, 303–312.

10 Collini, E. and Scholes, G.D. (2009)Coherent intrachain energy migration in aconjugated polymer at room temperature.Science, 323, 369–373.

11 Chan, F.T.S., Kaminski, C.F., andKaminski Schierle, G.S. (2010)HomoFRET fluorescence anisotropyimaging as a tool to study molecular self-assembly in live cells. Chemphyschem, 12,500–509.

12 van Ham, T.J., Esposito, A., Kumita, J.R.,Hsu, S.-T.D., Kaminski Schierle, G.S.,Kaminski, C.F., Dobson, C.M., Nollen,E.A.A., and Bertoncini, C.W. (2010)Towards multiparametric fluorescentimaging of amyloid formation: studies of aYFP model of [alpha]-synucleinaggregation. Journal of Molecular Biology,395, 627–642.

13 Jacobson, K., Mouritsen, O.G., andAnderson, R.G.W. (2007) Lipid rafts: at acrossroad between cell biology and physics.Nature Cell Biology, 9, 7–14.

14 Sharma, P., Varma, R., Sarasij, R.C., Ira,Gousset K., Krishnamoorthy, G., Rao, M.,and Mayor, S. (2004) Nanoscaleorganization of multiple GPI-anchoredproteins in living cell membranes. Cell,116, 577–589.

15 Varma, R. and Mayor, S. (1998) GPI-anchored proteins are organized insubmicron domains at the cell surface.Nature, 394, 798–801.

16 Bader, A.N., Hofman, E.G., Henegouwen,P., and Gerritsen, H.C. (2007) Imaging ofprotein cluster sizes by means of confocal

time-gated fluorescence anisotropymicroscopy. Optics Express, 15, 6934–6945.

17 Bader, A.N., Hofman, E.G., Voortman, J.,van Bergen en Henegouwen, P.M.P., andGerritsen, H.C. (2009) Homo-FRETimaging enables quantification of proteincluster sizes with subcellular resolution.Biophysical Journal, 97, 2613–2622.

18 Carlsson, C., Larsson, A., Jonsson, M.,Albinsson, B., and Nord�en, B. (1994)Optical and photophysical properties of theoxazole yellow DNA probes YO and YOYO.Journal of Physical Chemistry, 98, 10313–10321.

19 Carlsson, C., Larsson, A., Bj€orkman, A.,Jonsson, M., and Albinsson, B. (1997)Experimental and simulated fluorescencedepolarization due to energy transfer astools to study DNA–dye interactions.Biopolymers, 41, 481–494.

20 Larsson, A., Carlsson, C., Jonsson, M., andAlbinsson, B. (1994) Characterization ofthe binding of the fluorescent dyes YO andYOYO to DNA by polarized-lightspectroscopy. Journal of the AmericanChemical Society, 116, 8459–8465.

21 Hu, X.C., Damjanovic, A., Ritz, T., andSchulten, K. (1998) Architecture andmechanism of the light-harvestingapparatus of purple bacteria. Proceedings ofthe National Academy of Sciences of theUnited States of America, 95, 5935–5941.

22 Hu, X.C. and Schulten, K. (1997) Hownature harvests sunlight. Physics Today, 50,28–34.

23 Koepke, J., Hu, X.C., Muenke, C.,Schulten, K., and Michel, H. (1996) Thecrystal structure of the light-harvestingcomplex II (B800–850) from Rhodospirillummolischianum. Structure, 4, 581–597.

24 Ambroise, A., Kirmaier, C., Wagner, R.W.,Loewe, R.S., Bocian, D.F., Holten, D., andLindsey, J.S. (2002) Weakly coupledmolecular photonic wires: synthesis andexcited-state energy-transfer dynamics. TheJournal of Organic Chemistry, 67,3811–3826.

25 Wagner, R.W. and Lindsey, J.S. (1994) Amolecular photonic wire. Journal of theAmerican Chemical Society, 116, 9759–9760.

26 Wagner, R.W., Lindsey, J.S., Seth, J.,Palaniappan, V., and Bocian, D.F. (1996)Molecular optoelectronic gates. Journal of

References j649

the American Chemical Society, 118,3996–3997.

27 Lammi, R.K., Ambroise, A.,Balasubramanian, T., Wagner, R.W.,Bocian, D.F., Holten, D., and Lindsey, J.S.(2000) Structural control of photoinducedenergy transfer between adjacent anddistant sites in multiporphyrin arrays.Journal of the American Chemical Society,122, 7579–7591.

28 Hernando, J., Hoogenboom, J.P., van Dijk,E., Garcia-Lopez, J.J., Crego-Calama, M.,Reinhoudt, D.N., van Hulst, N.F., andGarcia-Parajo, M.F. (2004) Single moleculephotobleaching probes the exciton wavefunction in a multichromophoric system.Physical Review Letters, 93.

29 Tong, A.K., Jockusch, S., Li, Z.M., Zhu,H.R., Akins, D.L., Turro, N.J., and Ju, J.Y.(2001) Triple fluorescence energy transferin covalently trichromophore-labeled DNA.Journal of the American Chemical Society,123, 12923–12924.

30 Kawahara, S.-I., Uchimaru, T., and Murata,S. (1999) Sequential multistep energytransfer: enhancement of efficiency oflong-range fluorescence resonanceenergy transfer. Chemical Communications,563–564.

31 Heilemann, M., Tinnefeld, P., SanchezMosteiro, G., Garcia Parajo, M., Van Hulst,N.F., and Sauer, M. (2004) Multistepenergy transfer in single molecularphotonic wires. Journal of the AmericanChemical Society, 126, 6514–6515.

32 Tinnefeld, P., Heilemann, M., and Sauer,M. (2005) Design of molecular photonicwires based on multistep electronicexcitation transfer. Chemphyschem, 6,217–222.

33 Ohya, Y., Yabuki, K., Hashimoto, M.,Nakajima, A., and Ouchi, T. (2003)Multistep fluorescence resonanceenergy transfer in sequentialchromophore array constructed onoligo-DNA assemblies. BioconjugateChemistry, 14, 1057–1066.

34 Heilemann, M., Kasper, R., Tinnefeld, P.,and Sauer, M. (2006) Dissecting andreducing the heterogeneity of excited-stateenergy transport in DNA-based photonicwires. Journal of the American ChemicalSociety, 128, 16864–16875.

35 Sanchez-Mosteiro, G., van Dijk, E.M.,Hernando, J., Heilemann, M., Tinnefeld,P., Sauer, M., Koberlin, F., Patting, M.,Wahl, M., Erdmann, R. et al. (2006) DNA-based molecular wires: multiple emissionpathways of individual constructs. TheJournal of Physical Chemistry B, 110, 26349–26353.

36 Uemura, A., Kimura, S., and Imanishi, Y.(1983) Investigation on the interactions ofpeptides in the assembly of liposome andpeptide by fluorescence. Biochimica etBiophysica Acta, 729, 28–34.

37 Vyawahare, S., Eyal, S., Mathews, K.D., andQuake, S.R. (2004) Nanometer-scalefluorescence resonance optical waveguides.Nano Letters, 4, 1035–1039.

38 Hannestad, J.K., Sandin, P., andAlbinsson, B. (2008) Self-assembled DNAphotonic wire for long-range energytransfer. Journal of the American ChemicalSociety, 130, 15889–15895.

39 Su, W., Schuster, M., Bagshaw, C.R., Rant,U., and Burley, G.A. (2011) Site-specificassembly of DNA-based photonic wires byusing programmable polyamides.Angewandte Chemie, International Edition,50, 2712–2715.

40 Cohen, J.D., Sadowski, J.P., and Dervan,P.B. (2007) Addressing single molecules onDNA nanostructures. Angewandte Chemie,International Edition, 46, 7956–7959.

41 Cohen, J.D., Sadowski, J.P., and Dervan,P.B. (2008) Programming multiple proteinpatterns on a single DNA nanostructure.Journal of the American Chemical Society,130, 402–403.

42 B€orjesson, K., Preus, S., El-Sagheer, A.H.,Brown, T., Albinsson, B., andWilhelmsson, L.M. (2009) Nucleic acidbase analog FRET-pair facilitating detailedstructural measurements in nucleic acidcontaining systems. Journal of the AmericanChemical Society, 131, 4288–4293.

43 Iqbal, A., Arslan, S., Okumus, B., Wilson,T.J., Giraud, G., Norman, D.G., Ha, T., andLilley, D.M.J. (2008) Orientationdependence in fluorescent energy transferbetween Cy3 and Cy5 terminally attachedto double-stranded nuclelic acids.Proceedings of the National Academy ofSciences of the United States of America, 105,11176–11181.


44 Rothemund, P.W.K. (2006) Folding DNA tocreate nanoscale shapes and patterns.Nature, 440, 297–302.

45 Andersen, E.S., Dong, M., Nielsen, M.M.,Jahn, K., Subramani, R., Mamdouh, W.,Golas, M.M., Sander, B., Stark, H.,Oliveira, C.L.P. et al. (2009) Self-assemblyof a nanoscale DNA box with a controllablelid. Nature, 459, 73–76.

46 Douglas, S.M., Dietz, H., Liedl, T.,Hogberg, B., Graf, F., and Shih, W.M.(2009) Self-assembly of DNA intonanoscale three-dimensional shapes.Nature, 459, 414–418.

47 Gu, H.Z., Chao, J., Xiao, S.J., and Seeman,N.C. (2010) A proximity-basedprogrammable DNA nanoscale assemblyline. Nature, 465, 202–206.

48 Voigt, N.V., Torring, T., Rotaru, A.,Jacobsen, M.F., Ravnsbaek, J.B.,Subramani, R., Mamdouh, W., Kjems, J.,Mokhir, A., Besenbacher, F. et al. (2010)Single-molecule chemical reactions onDNA origami. Nature Nanotechnology, 5,200–203.

49 Stein, I.H., Steinhauer, C., and Tinnefeld,P. (2011) Single-molecule four-color FRETvisualizes energy-transfer paths on DNAorigami. Journal of the American ChemicalSociety, 133, 4193–4195.

50 Hannestad, J.K., Gerrard, S.R., Brown, T.,and Albinsson, B. (2011) Self-assembledDNA-based fluorescence waveguide withselectable output. Small, 7, 3178–3185.

51 Tumpane, J., Kumar, R., Lundberg, E.P.,Sandin, P., Gale, N., Nandhakumar, I.S.,Albinsson, B., Lincoln, P., Wilhelmsson,L.M., Brown, T. et al. (2007a) Triplexaddressability as a basis for functionalDNA nanostructures. Nano Letters, 7,3832–3839.

52 Tumpane, J., Sandin, P., Kumar, R.,Powers, V.E.C., Lundberg, E.P., Gale, N.,Baglioni, P., Lehn, J.M., Albinsson, B.,Lincoln, P. et al. (2007) Addressable high-information-density DNA nanostructures.Chemical Physics Letters, 440, 125–129.

53 Ruiz-Carretero, A., Janssen, P.G.A.,Stevens, A.L., Surin, M., Herz, L.M., andSchenning, A. (2011) Directing energytransfer in discrete one-dimensionaloligonucleotide-templated assemblies.Chemical Communications, 47, 884–886.

54 Benvin, A.L., Creeger, Y., Fisher, G.W.,Ballou, B., Waggoner, A.S., and Armitage,B.A. (2007) Fluorescent DNA nanotags:supramolecular fluorescent labels based onintercalating dye arrays assembled onnanostructured DNA templates. Journalof the American Chemical Society, 129,2025–2034.

55 €Ozhalici-€Unal, H. and Armitage, B.A.(2009) Fluorescent DNA nanotags basedon a self-assembled DNA tetrahedron. ACSNano, 3, 425–433.

56 Jullien, L., Canceill, J., Valeur, B., Bardez,E., Lef�evre, J.-P., Lehn, J.-M., Marchi-Artzner, V.r., and Pansu, R. (1996)Multichromophoric cyclodextrins. 4. Lightconversion by antenna effect. Journal of theAmerican Chemical Society, 118, 5432–5442.

57 Choi, M.-S., Aida, T., Yamazaki, T., andYamazaki, I. (2002) Dendriticmultiporphyrin arrays as light-harvestingantennae: effects of generation numberand morphology on intramolecular energytransfer. Chemistry, 8, 2667–2678.

58 Choi, M.-S., Yamazaki, T., Yamazaki, I.,and Aida, T. (2004) Bioinspired moleculardesign of light-harvesting multiporphyrinarrays. Angewandte Chemie, InternationalEdition, 43, 150–158.

59 Aratani, N., Kim, D., and Osuka, A. (2009)Discrete cyclic porphyrin arrays as artificiallight-harvesting antenna. Accounts ofChemical Research, 42, 1922–1934.

60 Hoffmann, M., Karnbratt, J., Chang, M.H.,Herz, L.M., Albinsson, B., and Anderson,H.L. (2008) Enhanced pi conjugationaround a porphyrin[6] nanoring.Angewandte Chemie, International Edition,47, 4993–4996.

61 Nakamura, Y., Aratani, N., and Osuka, A.(2007) Cyclic porphyrin arrays as artificialphotosynthetic antenna: synthesis andexcitation energy transfer. Chemical SocietyReviews, 36, 831–845.

62 Schenning, A.P.H.J., Benneker, F.B.G.,Geurts, H.P.M., Liu, X.Y., and Nolte, R.J.M.(1996) Porphyrin wheels. Journal of theAmerican Chemical Society, 118, 8549–8552.

63 Yang, J., Park, M., Yoon, Z.S., Hori, T.,Peng, X.B., Aratani, N., Dedecker, P.,Hotta, J.I., Uji-I, H., Sliwa, M. et al. (2008)Excitation energy migration processes incyclic porphyrin arrays probed by single

References j651

molecule spectroscopy. Journal of theAmerican Chemical Society, 130, 1879–1884.

64 Sautter, A., KaletaSs, B.K., Schmid, D.G.,Dobrawa, R., Zimine, M., Jung, G., vanStokkum, I.H.M., De Cola, L., Williams,R., and W€urthner, F. (2005) Ultrafastenergy-electron transfer cascade in amultichromophoric light-harvestingmolecular square. Journal of the AmericanChemical Society, 127, 6719–6729.

65 Wurthner, F. and Sautter, A. (2003) Energytransfer in multichromophoric self-assembled molecular squares. Organic andBiomolecular Chemistry, 1, 240–243.

66 Goodman, R.P., Berry, R.M., andTurberfield, A.J. (2004) The single-stepsynthesis of a DNA tetrahedron. ChemicalCommunications, 1372–1373.

67 Miller, R.A., Presley, A.D., and Francis, M.B. (2007) Self-assembling light-harvestingsystems from synthetically modifiedtobacco mosaic virus coat proteins. Journalof the American Chemical Society, 129,3104–3109.

68 Brousmiche, D.W., Serin, J.M., Fr�echet,J.M.J., He, G.S., Lin, T.C., Chung, S.J.,Prasad, P.N., Kannan, R., and Tan, L.S.(2004) Fluorescence resonance energytransfer in novel multiphoton absorbingdendritic structures. The Journal of PhysicalChemistry B, 108, 8592–8600.

69 Dutta, P.K., Varghese, R., Nangreave, J.,Lin, S., Yan, H., and Liu, Y. (2011) DNA-directed artificial light-harvesting antenna.Journal of the American Chemical Society,133, 11985–11993.

70 Wang, R., Liu, W., and Seeman, N.C.(2009) Prototyping nanorod control: a DNAdouble helix sheathed within a DNA six-helix bundle. Chemistry & Biology, 16,862–867.

71 Asanuma, H., Ito, T., Yoshida, T., Liang,X.G., and Komiyama, M. (1999)Photoregulation of the formation anddissociation of a DNA duplex by using thecis–trans isomerization of azobenzene.Angewandte Chemie, International Edition,38, 2393–2395.

72 Asanuma, H., Takarada, T., Yoshida, T.,Tamaru, D., Liang, X.G., and Komiyama,M. (2001) Enantioselective incorporation ofazobenzenes into oligodeoxyribonucleotidefor effective photoregulation of duplex

formation. Angewandte Chemie,International Edition, 40, 2671–2673.

73 Liang, X.G., Asanuma, H., and Komiyama,M. (2002) Photoregulation of DNAtriplex formation by azobenzene. Journalof the American Chemical Society, 124,1877–1883.

74 Andersson, J., Li, S.M., Lincoln, P., andAndreasson, J. (2008) Photoswitched DNA-binding of a photochromic spiropyran.Journal of the American Chemical Society,130, 11836–11837.

75 Sapsford, K.E., Berti, L., and Medintz, I.L.(2006) Materials for fluorescenceresonance energy transfer analysis: beyondtraditional donor–acceptor combinations.Angewandte Chemie, International Edition,45, 4562–4588.

76 Alivisatos, A.P., Gu, W.W., and Larabell, C.(2005) Quantum dots as cellular probes.Annual Review of Biomedical Engineering,755–76.

77 Klostranec, J.M. and Chan, W.C.W. (2006)Quantum dots in biological and biomedicalresearch: recent progress and presentchallenges. Advanced Materials, 18, 1953–1964.

78 Michalet, X., Pinaud, F.F., Bentolila, L.A.,Tsay, J.M., Doose, S., Li, J.J., Sundaresan,G., Wu, A.M., Gambhir, S.S., and Weiss, S.(2005) Quantum dots for live cells, in vivoimaging, and diagnostics. Science, 307,538–544.

79 Boeneman, K., Deschamps, J.R.,Buckhout-White, S., Prasuhn, D.E.,Blanco-Canosa, J.B., Dawson, P.E., Stewart,M.H., Susumu, K., Goldman, E.R.,Ancona, M. et al. (2010) Quantum dotDNA bioconjugates: attachment chemistrystrongly influences the resulting compositearchitecture. ACS Nano, 4, 7253–7266.

80 Boeneman, K., Prasuhn, D.E., Blanco-Canosa, J.B., Dawson, P.E., Melinger, J.S.,Ancona, M., Stewart, M.H., Susumu, K.,Huston, A., and Medintz, I.L. (2010) Self-assembled quantum dot-sensitizedmultivalent DNA photonic wires.Journal of the American Chemical Society,132, 18177–18190.

81 B€orjesson, K., Tumpane, J., Ljungdahl, T.,Wilhelmsson, L.M., Nord�en, B., Brown, T.,Ma

�rtensson, J., and Albinsson, B. (2009)

Membrane-anchored DNA assembly for


energy and electron transfer. Journal of theAmerican Chemical Society, 131, 2831–2839.

82 B€orjesson, K., Wiberg, J., El-Sagheer, A.H.,Ljungdahl, T., Ma

�rtensson, J., Brown, T.,

Nord�en, B., and Albinsson, B. (2010)Functionalized nanostructures: redox-active porphyrin anchors forsupramolecular DNA assemblies. ACSNano, 4, 5037–5046.

83 B€orjesson, K., Lundberg, E.P., Woller, J.G.,Nord�en, B., and Albinsson, B. (2011) Soft-surface DNA nanotechnology: DNAconstructs anchored and aligned to lipidmembrane. Angewandte Chemie,International Edition, 50, 8312–8315.

84 B€orjesson, K., Woller, J.G., Parsa, E.,Ma

�rtensson, J., and Albinsson, B. (2011) A

bioinspired self assembled dimericporphyrin pocket that binds electronaccepting ligands. ChemicalCommunications, 48, 1793–1795.

85 D’Souza, F., Smith, P.M., Zandler, M.E.,McCarty, A.L., Itou, M., Araki, Y., and Ito,O. (2004) Energy transfer followed byelectron transfer in a supramolecular triadcomposed of boron dipyrrin, zincporphyrin, and fullerene: a model for thephotosynthetic antenna–reaction centercomplex. Journal of the American ChemicalSociety, 126, 7898–7907.

86 Gust, D., Moore, T.A., and Moore, A.L.(1993) Molecular mimicry of

photosynthetic energy and electrontransfer. Accounts of Chemical Research, 26,198–205.

87 Gust, D., Moore, T.A., and Moore, A.L.(2001) Mimicking photosynthetic solarenergy transduction. Accounts of ChemicalResearch, 34, 40–48.

88 Kuciauskas, D., Liddell, P.A., Lin, S.,Johnson, T.E., Weghorn, S.J., Lindsey, J.S.,Moore, A.L., Moore, T.A., and Gust, D.(1999) An artificial photosyntheticantenna–reaction center complex. Journalof the American Chemical Society, 121,8604–8614.

89 Moore, T.A., Gust, D., Mathis, P., Mialocq,J.C., Chachaty, C., Bensasson, R.V., Land,E.J., Doizi, D., Liddell, P.A., Lehman, W.R.et al. (1984) Photodriven charge separationin a carotenoporphyrin–quinone triad.Nature, 307, 630–632.

90 Steinberg-Yfrach, G., Liddell, P.A., Hung,S.C., Moore, A.L., Gust, D., and Moore, T.A. (1997) Conversion of light energy toproton potential in liposomes by artificialphotosynthetic reaction centres. Nature,385, 239–241.

91 Woller, J.G., Hannestad, J.K., andAlbinsson, B. (2013) Self-assemblednanoscale DNA–porphyrin complexfor artificial light harvesting.Journal of the American Chemical Society,135, 2759–2768.

References j653

Date post:	23-Dec-2016
Category:	Documents
Upload:	niko
View:	220 times
Download:	6 times

FRET - Förster Resonance Energy Transfer || Multistep FRET and Nanotechnology

Documents