A coordination polymer for the site-specific integration of semiconducting sequences into DNA-based materials

Advances in bottom-up material design have been significantly progressed through DNA-based approaches. However, the routine integration of semiconducting properties, particularly long-range electrical conduction, into the basic topological motif of DNA remains challenging. Here, we demonstrate this with a coordination polymer derived from 6-thioguanosine (6-TG-H), a sulfur-containing analog of a natural nucleoside. The complexation reaction with Au(I) ions spontaneously assembles luminescent one-dimensional helical chains, characterized as {AuI(μ-6-TG)}n, extending many μm in length that are structurally analogous to natural DNA. Uniquely, for such a material, this gold-thiolate can be transformed into a wire-like conducting form by oxidative doping. We also show that this self-assembly reaction is compatible with a 6-TG-modified DNA duplex and provides a straightforward method by which to integrate semiconducting sequences, site-specifically, into the framework of DNA materials, transforming their properties in a fundamental and technologically useful manner.

Supplementary Discussion FTIR spectroscopy. The FTIR spectrum of 6-TGH itself shows an intense v(C=S) stretch at 1205 cm -1 and the lack of absorbance corresponding to v(S-H), around 2600 cm -1 , indicating the parent compound is in the thione-, rather than thiol-, tautomeric form. 1,2 Upon reaction with Au(I) ions this band disappears indicating metal ion binding at the S6 group in the thiolform, consistent with literature for Au(I)-6-thiopurine derivatives ( Supplementary Fig. 2). [3][4][5][6] Electrospray Mass Spectrometery. Electrospray mass spectrometry data of samples of 1 confirmed the presence of oligomers consistent with the formation of a coordination polymer.  suggest some partial ordering of 1 on the substrate. The best-fit diameter is less than that observed in AFM, but consistent with the proposed structure of 1 in which the electron density is concentrated in the Au-S chain in the interior of the helix.

Molecular modeling and DFT calculations.
A model for 1 consistent with the above data is shown in Fig. 2 as a {Au 8 (µS-6TG) 8 } oligomeric chain. This is based on the single-crystal Xray diffraction study of the analogous gold(I)-thiolate, derived from thiomalate. 11 This structural motif has also recently been observed for gold(I)-thiophenolate. 7 The model of 1 is a helical 1-D coordination polymer with a central {-µS6-Au-} n chain, about which the nucleosides are pendant. The featured 2-S6 bridging mode is known for 6-thiopurine derivatives in forming both discrete 12,13 and polymeric coordination complexes. [14][15][16] The diameter of the helix in 1 at approx. 2.0 nm, as measured between hydroxyl-group oxygen atoms, is consistent with AFM data.
The DFT calculations used Firefly QC package 17 , which is partially based on the GAMESS (US) source code, 18 with the SBKJC effective core potentials. 19 Prior to treatment with iodine the component at 84.6 eV is attributed to Au 4f 7/2 and that at 88.3 eV to Au 4f 5/2 . 21,22 The BE value of the 4f 7/2 peak indicates the +1 formal oxidation state for the Au ions 23 (typical range 84.3-85.5 eV). 24,25 After doping with iodine, both the Au 4f 5/2 and 4f 7/2 peaks are shifted to higher binding energy by 0.2 eV.
The high resolution XPS spectra for sulfur were fitted with two spin-orbit components (2p 1/2 and 2p 3/2 ) with a fixed splitting ~1.2 eV and with the ratio between the band areas fixed at 1:2. The S 2p 3/2 at BE 163 eV is attributed to sulfur in Au(I)-thiolate form. 22 This value, in fact, suggests a bridging thiolate mode in 1 based on previous reports which show higher S 2p 3/2 binding energy (~163 eV) 26,27 for this compared to comparable terminal thiol binding (~162 eV) 28  Additionally, the S 2p 3/2 spectra is shifted after oxidation to a higher BE by 0.2-0.3 eV. The high resolution scan of XPS spectra for iodine were fitted with two spin-orbit components (3d 3/2 and 3d 5/2 ) separated by 11.5 eV and with a component band area ratio of 2:3. 22 The I The M/S/M system is symmetric and therefore the overall resistance is: As long as ≫ , then most of the applied bias is dropped across the bulk of S and the I-V characteristic is ohmic irrespective of whether S is a metal or semiconductor. It is not easy to distinguish semiconducting from metallic behavior of the film based on the I-V characteristic alone. However, the temperature dependence of the measured resistance R allows us to make the distinction because metallic conductors show an increased resistance as the temperature rises due to increased electron-phonon scattering rates and semiconductor crystals or polymers show a decreased resistance because of the effect of thermal activation on either carrier density or hopping rates (Fig. 3h, main text).
At sufficiently high bias, it may be possible to observe deviations from ohmic behavior (see undoped polymer in Figure 3g of the manuscript and discussion below). We and others have seen such effects in other polymer systems, 44-46 but we do not observe them in the present work for the doped polymer which shows ohmic behaviour consistent with supplementary equation (3). However, it can be seen that in the undoped polymer, potentials > 1.5 V produce a deviation from linearity, although the trace is still reversible on the return scan.
This part of the I-V curve is similar to those observed for hopping polymers under conditions where the ionic countercharges are immobile. 44 It has a similar functional form to the Butler-Volmer equation.
n and n p are the densities of empty and occupied sites (by carriers),  is the site-site hopping distance, L is the interelectrode spacing and k ex is the electron self-exchange rate constant for site-site hopping at zero bias. D is the dimensionality of the system,  is a symmetry factor (about 0.5 in Marcus theory), and  is a non-ideality factor that is often interpreted in terms of clustering of sites. In the case of the undoped Au-6TG polymer n p is very small and there is little current until charge is injected at the electrodes. It is worth noting that supplementary equation (4) predicts the linear region at low bias and an Arrhenius-like temperature dependence.
Electrochemical effects and the DNA bases. We considered whether the I-V measurements could be interpreted in terms of electrochemical oxidation of the bases. However, the linearity and lack of hysteresis of the I-V data in Fig. 3g of the manuscript strongly argues against this. The lack of hysteresis (doped or undoped polymer) and the linear I-V characteristic (doped polymer) are inconsistent with an electrochemical process driven by the interfacial potential and especially with the electrochemistry of the DNA bases where chemical steps following the electron transfer prevent observation of the reverse process except at scan rates (10 -100 V s -1 ) orders of magnitude greater than applied in our work. 47 Instead, the lack of hysteresis is a piece of evidence in favour of our interpretation of the measurements in terms of electron transport in the film.