Nanopore-Based SensorsDOI: 10.1002/anie.201200688

DNA Origami Gatekeepers for Solid-State Nanopores**Ruoshan Wei, Thomas G. Martin, Ulrich Rant,* and Hendrik Dietz*

Molecular self-assembly with DNA enables the constructionof soluble objects with nanometer to micrometer scaleabsolute dimensions and custom chemical features,[1] includ-ing crystals,[2] patterns,[3] bricks,[4] boxes,[5] and curvedshapes,[6] that can open novel paths to scientific discovery.[7]

Herein, we report on DNA nanoplates for nanopore-basedsensing approaches. Nanopores in biological or solid-statemembranes offer great potential for label-free single-mole-cule sensing applications.[8] Biological nanopores, such asalpha-hemolysin, can be customized within the limits ofprotein engineering.[9] Artificial nanopores in solid-statemembranes can be made with user-defined dimensions, butchemical modifications require substantial effort.[10] A chal-lenge in the field is to gain control over both the geometricaland chemical specifications of nanopores. We hypothesizedthat using DNA-based nanoplates as covers for solid-statenanopores could provide a route for meeting this challenge.

Our setup (Figure 1 a) consists of two electrolyte reser-voirs separated by a silicon-supported free-standing insulatingsilicon nitride (SiN) membrane of thickness L = 50 nm. Themembrane contains a single conical nanopore of diameterD = 1825 nm (Figure 1b), which is fabricated by electronbeam lithography and reactive ion etching.[11] When a voltageis applied through the two electrodes, an ionic current flowsthrough the nanopore and is recorded with a current ampli-fier. The dimensions and shape of the nanopore dominate theresistance of the setup. The cis side of the nanopore is coveredwith a rectangular nanoplate of width a, length b, andthickness l. The plate includes a central aperture of widthx and length y (Figure 1a). The nanoplates are produced bymolecular self-assembly with scaffolded DNA origami (see

the Supporting Information, note S1) and consist of a doublelayer of 46 tightly interlinked double-helical DNA domains ina honeycomb-type packing lattice. We made nanoplateversions (Supporting Information, Figures S1S5) witha width and length of 50 nm, and a thickness of 6 nm. Correctformation of the nanoplates was confirmed using negative-stain transmission electron microscopy (Figure 1b; see alsothe Supporting Information, notes S6S9).

The nanoplates were electrically assembled onto thenanopores by injection into the cis electrolyte compartment.A sudden current drop and intensification in the current noisewas typically observed within a few seconds (Figure 1c) afterthe bias voltage was applied (Supporting Information,notes S2 and S3). The current blockades lasted for hours,unless reverse voltages were applied or the membrane wasintensely rinsed. In both cases, the initial conductance level ofthe nanopore was restored. In experiments using nanoporeswith diameters exceeding the dimensions of the nanoplates,we observed transient blockades (Supporting Information,note S4), indicating that in those cases the nanoplates slippedthrough the larger nanopores. In experiments with mem-branes containing arrays of nanopores, we observed staircase-like decreases in conductivity (Supporting Information, Fig-ure S12), indicating the progressive capture of nanoplates byindividual nanopores in the array.

We find that, for nanoplates with apertures the conduc-tance of nanoplate-on-nanopore hybrids decreases withdecreasing aperture size (Figure 1d). This finding can beexplained by a model in which the nanoplates cover thenanopore in a flat orientation. Orthogonal or randomorientations should yield relative conductances that do notdepend on the size of the apertures. The nanopore conduc-tance drops by approximately 17% when using nanoplateslacking a central aperture. The DNA nanoplates are expectedto be permeable to ions, owing to a mesh of vertical andhorizontal channels. The vertical channels are formed bycavities between neighboring double-helical DNA domainsbetween crosslinks along the helical axis,[4, 5,12] while thehorizontal channels are intrinsic to the honeycomb-typearchitecture of the DNA nanoplate. The corrugated lowersurface of a nanoplate may provide additional channels forion flow when it covers a nanopore.

The equivalent circuit model depicted in Figure 1e, faccounts for the conductances of the bare nanopore (Gpore),the nanoplate (Gplate), and the central aperture (Gaperture).Gaperture can be estimated for a given aperture size bya geometrical conductance model (Supporting Information,note S5). The value of Gplate can therefore be computed fromthe measured hybrid conductance. We evaluated Gplate versusGpore for nanopores with diameters in the range of D = 1825 nm and found that the conductance Gplate of the nanoplatesincreases linearly with the conductance Gpore of the nanopore

[*] R. Wei,[+] U. RantWalter Schottky Institute, Technische Universitt MnchenAm Coulombwall 4, 85748 Garching near Munich (Germany)E-mail: rant@tum.de

T. G. Martin,[+] H. DietzCenter for Integrated Protein Science Mnchen & Institute forAdvanced Study, Physics Department, Technische UniversittMnchenAm Coulombwall 4a, 85748 Garching near Munich (Germany)E-mail: dietz@tum.de

[+] These authors contributed equally to this work.

[**] This work was supported by the German Excellence Initiativethrough grants from the Nano Initiative Munich and from theCenter for Integrated Protein Science Munich, by the CollaborativeResearch Center SFB 863 of the German Research Society (DFG),the TUM Institute for Advanced Study, and a European ResearchCouncil Starting Grant to HD. R.W. was supported by the TUMGraduate Schools Faculty Graduate Center of Physics. Discussionswith Friedrich Simmel, Michael Mayer, Daniel Branton, and GeorgeChurch are gratefully acknowledged.

Supporting information for this article is available on the WWWunder http://dx.doi.org/10.1002/anie.201200688.

.AngewandteCommunications

4864 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2012, 51, 4864 4867

http://dx.doi.org/10.1002/anie.201200688

(Figure 1g). Therefore, thearea of the nanopore con-trols the area of the nano-plate that is exposed to ionflow. This finding agrees wellwith a geometric model(Supporting Information,note S5) that further sup-ports the notion that thenanoplates cover the nano-pore in the desired orienta-tion. Neglecting potentialion fluxes that bypass thenanoplates, we estimate thetransversal specific conduc-tivity (splate?) of the nano-plates to be (7.8 0.4) Sm1in an aqueous solution ofKCl (1m) and MgCl2(10 mm).

DNA nanoplates withcustom apertures can serveas size-selective molecular

gates for nanopores. Todemonstrate this option,we added streptavidin(52 kDa, hydrodynamicdiameter ca. 6 nm)[13] tothe cis compartment of thenanopore setup along withnanoplates containinga central aperture ofwidth 9 nm and length14 nm. Before insertionof a nanoplate onto thenanopore, we observedsparse sub-millisecondcurrent blockades causedby streptavidin moleculestranslocating through thebare nanopore (Fig-ure 2b). After nanoplateinsertion, the currentblockades caused by strep-tavidin translocationthrough the nanoplate-on-nanopore persisted(Figure 2b,c; see also theSupporting Information,note S6). When werepeated these experi-ments with a significantlylarger protein, immuno-globulin G (IgG, 150 kDa,effective hydrodynamicdiameter ca. 14 nm), wefound that with a nano-plate in place, IgG trans-location was inhibited

Figure 1. DNA nanoplates on SiN nanopores. a) Schematic of the experimental setup. The free-standing SiNmembrane is shown in grey, the DNA nanoplate is shown in red. Inspired by previous work,[10d] we includeda ca. 1300 base long single-stranded or double-stranded DNA loop that protrudes near the central aperture (seeFigures S1-S5 for design details) to facilitate insertion into the nanopore. b) Top left: Transmission electronmicrograph (TEM) of a bare nanopore. Top right and bottom row: average of aligned negative-stain TEMmicrographs of nanoplates with varying aperture sizes (inset numbers indicate aperture width length in nm)imaged separately on thin carbon support layers (Figures S6S9). Scale bars: 20 nm. c) Nanoplates wereinjected at dilute effective concentrations of ca. 30 pmolL1 with bias voltages of up to + 200 mV applied to thetrans electrode and captured electrically on nanopores. Horizontal scale bars: 2 s; Vertical scale bars: 10 nS.Asterisk marks nanoplate capture event. d) Relative conductance of nanplate-on-nanopore hybrids compared toconductance of bare nanopore before nanoplate insertion. Inset numbers indicate width length in nm ofa central aperture in the nanoplates; 5 7 indicates a nanoplate version with one single-stranded DNAheptanucleotide overhang in the aperture (Figure S2). e, f) Schematic current paths and equivalent circuit modelfor the conductance of the nanoplate-on-nanopore hybrid. g) Nanoplate conductance Gplate correlates linearlywith the conductance of the underlying nanopore Gpore (note S5); & no aperture, * 5 nm 7 nm aperture, and ~9 nm 14 nm aperture.

Figure 2. DNA nanoplates with custom apertures for size-selective macromolecular sensing with SiNnanopores. a) Currenttime traces observed for streptavidin (cis concentration= 20 nm) translocation throughbare nanopores, b) Currenttime traces after insertion of a nanoplate with a central aperture of width 9 nmand length 14 nm (see Figure S5 for design details). c) Histogram of observed current levels. df) As in (a)(c), but with Immunoglobulin G (IgG; cis concentration = 20 nm). gi) As in (a)(c), but with 6 kbp lineardouble-helical DNA molecules (cis concentration = 300 pm) and a nanoplate with aperture width 5 nm andlength 7 nm (Figure S2, version 0).

AngewandteChemi