Insulating Behavior of $\lambda$-DNA on the Micron Scale

We have investigated the electrical conductivity of $\lambda$-DNA using DNA covalently bonded to Au electrodes. Thiol-modified dTTP was incorporated into the “sticky” ends of bacteriophage $\lambda$-DNA using DNA polymerase. Two-probe measurements on such molecules provide a hard lower bound for the resistivity $\rho >{10}^6\Omega \mathrm{c}\mathrm{m}$ at bias potentials up to 20 V, in conflict with recent claims of moderate to high conductivity. By direct imaging, we show that the molecules are present after the measurements. We stress the importance of eliminating salt residues in these measurements.

Figure 1

(A) The schematic of the incorporation of deoxynucleoside triphosphates into $\lambda$-DNA sticky ends using the Klenow fragment ($3^{\prime }\to 5^{\prime }$ ${\mathrm{e}\mathrm{x}\mathrm{o}}^{-}$); (B) pulsed-field gel (PFG) electrophoresis of thiol-modified and natural $\lambda$-DNA before and after ligation [18]. Lane 1: $\lambda$-DNA PFG marker (New England Biolabs); 2: unmodified natural-$\lambda$ monomers; 3: ligated natural-$\lambda$ multimers; 4: thiol-modified $\lambda$-DNA (HS-$\lambda$) monomers; 5: HS-$\lambda$ after ligation, with no significant multimer formation; 6: control-$\lambda$ after ligation, proving that the presence of unincorporated ${\mathrm{S}}^4$-dTTP and other dNTPs does not inhibit the ligation reaction.

Figure 2

Images of HS-$\lambda$ DNA on quartz chips with Au electrodes. The DNA molecules were prestained with TOTO1. Observations were through a Nikon Eclipse TE300 inverted microscope equipped with a $60\times$ oil-immersion lens ($\mathrm{N}\mathrm{A}=1.4$) and excited by a collimated Ar:Kr ion laser at 488 nm. Fluorescence images were collected at 533 nm by an intensified CCD camera (Princeton Instruments) and digitally enhanced. Au electrodes appear as vertical stripes. Arrows represent the flow direction. Panel A1 and A2: Molecule A, anchored at both ends, and flexing with the flow. Panel B1 and B2: Both ends of molecule B were attached on the same electrode, with the midsegment moving freely in the buffer, showing that the attachment was specific to the thiol-modified ends. Panel C: Many HS-$\lambda$ DNA molecules spanning two electrodes. The image was taken after the ${\mathrm{M}\mathrm{g}}^{2+}$ and ${\mathrm{N}\mathrm{H}}_4\mathrm{A}\mathrm{c}$ rinsing. Scale bar applies to all panels [24].

Figure 3

The two-probe current vs voltage ($I\mathrm{\text{−}}V$) curve for a sample of $\lambda$-DNA bridging two parallel Au electrodes separated by $4\mu \mathrm{m}$ (the sample comprises $\sim 1000$ molecules). The DNA was rinsed with ${\mathrm{N}\mathrm{H}}_4\mathrm{A}\mathrm{c}$ before the measurement to remove the buffer salt residue. The dashed line is a linear fit to the data. The inset shows the two-probe $I\mathrm{\text{−}}V$ curve for a test chip containing $1\times \mathrm{T}\mathrm{E}$ buffer solution (without DNA). The chip was dried in vacuum but not subject to ${\mathrm{N}\mathrm{H}}_4\mathrm{A}\mathrm{c}$ rinsing. The observed conductance is entirely from trace TE salt residue. Both measurements were done in vacuum ($<{10}^{-7}\mathrm{T}\mathrm{o}\mathrm{r}\mathrm{r}$) at 295 K. Open-circuit impedance between any two electrodes was always $\gg 10\mathrm{T}\mathrm{\Omega }$.