A modular LHC built on the DNA three-way junction †

A light-harvesting complex composed of a p -stacked multichromo-phoric array in a DNA three-way junction is described. The modular design allows for a ready exchange of non-covalently attached energy acceptors.

The architecture of the 3WJ, its sequences and the corresponding building blocks are shown in Scheme 1. The phenanthrene (P), 52 pyrene (S and X) 53,54 perylenediimide (E), 55 and the commercially available cyanine (Cy) phosphoramidite building blocks were incorporated into the oligomers by automated oligonucleotide synthesis. All oligomers were purified by reverse phase HPLC and analysed by LC-MS (see ESI †). In 3WJ-2, -3, -4 and -5, the LHC is composed of a phenanthrene (P) stack, the light absorbing antenna (strands a and b), and an exchangeable acceptor chromophore (strands d-g). 3WJ-1 serves as a control bearing no acceptor dye in the third strand (c). Formation of the 3WJs was demonstrated by polyacrylamide gel electrophoresis (ESI †). Thermal denaturation experiments showed a single transition for all 3WJs. Melting temperature (T m ) values ranging from 57-63 1C are listed in Table 1. UV/Vis spectra of the different 3WJs ( Fig. 1) reveal the following long-wavelength absorptions of the various chromophores: P (l max = 316 nm); S (l max = 353 nm); X (l max = 391 nm); E (l max = 561 nm); and Cy (l max = 652 nm).
The effectiveness of the 3WJs as LHCs was investigated by fluorescence spectroscopy. Fig. 2 shows the emission spectra of 3WJ-1 to 3WJ-4 after excitation of phenanthrene at 320 nm. 3WJ-2 and 3WJ-3, both of which contain a pyrene derivative as acceptor, exhibit strong Scheme 1 Schematic illustration of the light harvesting 3WJ, the chromophores used and the corresponding sequences.  exciplex emission (l max = 435 and 455 nm, respectively 56 ). This is in agreement with the observation made previously with a duplex containing a phenanthrene antenna with a covalently attached pyrene (S). 46 In the present case, however, the acceptor dye is brought into proximity of the phenanthrene antenna in a non-covalent fashion. A quantum yield of 0.21 (Table 1) is obtained for 3WJ-2. Similar results are obtained for 3WJ-3. However, some pyrene monomer fluorescence (peak at 400 nm and shoulder at around 425 nm) is present in addition to the exciplex emission. The total quantum yield of 3WJ-3 with 1,6-dialkynylpyrene as the acceptor dye is 0.32.
Irradiation of the control 3WJ-1, which contains the phenanthrene stack but no acceptor chromophore, at 320 nm results in a broad, unstructured emission band centred around 415 nm, which is assigned to phenanthrene excimer fluorescence. 57-60 F F for 3WJ-1 was 0.10 (Table 1). This is considerably less than the values found for 3WJ-2 and -3 (0.21 and 0.32) that contain a terminal pyrene chromophore.
In three-way junction 3WJ-4, a perylenediimide unit is used in strand c. This chromophore quenches the phenanthrene excimer fluorescence to a considerable extent (F F 0.02). The quenching effect 61,62 is strong, but not as pronounced as in the case of pyrene excimer quenching by the same perylenediimide derivative. 37,63 The reduced effect can be rationalized by a smaller spectral overlap in the present case (phenanthrene-perylenediimide) than between pyrene and perylenediimide. 63 The emission spectra obtained by direct excitation of the twopyrene acceptor dyes (S or X) are shown in Fig. 3. Excitation was carried out at 370 nm, at which wavelength both pyrene derivatives absorb and phenanthrene excitation is negligible. Two aspects are worth noting: (i) the fluorescence intensities are significantly lower than that obtained after excitation of phenanthrene, which shows that the phenanthrene-pyrene 3WJ functions as a LHC; and (ii) the fraction of monomer emission in 3WJ-3 is somewhat larger than after phenanthrene excitation (cf. Fig. 2). Furthermore, direct excitation of control 3WJ-1 at 370 nm (Fig. 3) results in no significant fluorescence.
Energy transfer is also observed from phenanthrenes to Cy. Fig. 4 shows the effect of a stepwise build-up of 3WJ-5. In this experiment, the phenanthrene antenna (i.e. a 1:1-mixture of strands a and b) is added stepwise to the acceptor dye (strand g). A gradual increase in Cy fluorescence is observed until antenna and acceptor are present in equimolar concentrations. Phenanthrene excimer emission is also present and grows parallel to Cy fluorescence. Further addition of the antenna beyond the 1 : 1-ratio does not result in additional growth of Cy fluorescence, while phenanthrene fluorescence grows proportionally. The quantum yields for P and Cy emission amount to 0.015 and 0.049, respectively (Table 1). Cy has no significant absorbance (see ESI, † Fig. S28) at the irradiation wavelength (320 nm) or in the region of phenanthrene emission (360-530 nm). As shown by the corresponding excitation spectra (Fig. 5), phenanthrene excitation (320 nm) leads to Cy emission only in 3WJ-5 but not in single strand g. These experiments demonstrate that Cy fluorescence in 3WJ-5 is due to energy transfer from the phenanthrenes. 64 A Förster radius of B26 Å is calculated (ESI †). Considering the dimensions of the cavity in a typical 3WJ (B12 Å) 65 as well as its inherent flexibility, 33,65 energy transfer via FRET is, thus, possible. On the other hand, molecular contact between the donor and the acceptor is also     feasible (as e.g. shown by exciplex formation in 3WJ-2 and -3) rendering excitation transfer via Dexter electron exchange a plausible mechanism. 61,66 In conclusion, we have described the construction of LHCs on the basis of the DNA 3WJ. The LHC is based on a modular construction in which a phenanthrene antenna is located in one of the three stems and the acceptor is brought into proximity of the antenna through the annealing of the third strand. Phenanthrene excitation (320 nm) is followed by energy transfer to pyrene, perylenediimide or a cyanine dye. In the case of pyrene, exciplex fluorescence is observed, whereas perylenediimide acts as a quencher. In the case of cyanine, energy transfer leads to cyanine fluorescence. The study demonstrates the use of the DNA 3WJ for the modular assembly of artificial LHCs. This work was supported by the Swiss National Foundation (Grant 200020-149148).