IONIC CURRENT DETECTION OF DNA ORIGAMI NANOSTRUCTURES WITH NANOCAPILLARIES

N. A. W. Bell, S. M. Hernández-Ainsa, C. R. Engst, T. Liedl, and U. F. Keyser

1 INTRODUCTION

A continuing challenge in the nanopore field is the control of geometry and surface functionality on the nanometre scale. Protein nanopores such as α-haemolysin can be genetically engineered to attach binding motifs at specific positions but the narrow channel size limits most membrane proteins to the study of single stranded DNA and unfolded proteins. The mechanical instabilities of lipid bilayers and controlling the number of protein insertions also create difficulties for device integration. Synthetic nanopores made by silicon nanotechnology can be tuned to different diameters for the analysis of a wide range of biomolecules but the creation of fixed structures with atomic scale control has yet to be demonstrated. Hybrid nanopores offer a novel method for the formation of nanopores by directing a molecular construct into a silicon based nanopore. This combines the advantages of a solid membrane support with a fixed number of pores and the ability to make atomically defined molecular constructs made by self-assembly.

We recently demonstrated first experiments on forming a hybrid DNA origami nanopore. DNA origami is a well established technique for the formation of nanostructures with almost arbitrary geometry. A 7-8kbp single strand of DNA is folded into a pre-programmed structure by adding 100s of short synthesised DNA strands known as 'staple' strands. Functional motifs can be added by standard biochemical modifications to the staple strands. The process of design and synthesis of a DNA origami structure can be carried out in less than a week and several billion structures made simultaneously in high yields. This makes DNA origami ideally suited for the rapid prototyping of designer nanopore structures with the potential for massive parallelisation.

Here we describe the fabrication of a flat DNA origami nanopore structure. Our experiments show that translocations and trapping of these structures can be readily detected with nanocapillaries.

2 METHOD AND RESULTS

2.1 Fabrication of nanocapillaries

The fabrication of nanocapillaries and their use for detecting the folding state of DNA has been described in detail elsewhere. Briefly, a laser assisted pipette puller (Sutter P-2000) is used to draw down the diameter of quartz capillaries to a few tens of nanometres. The nanocapillary is then glued into a PDMS mold so that it is the only connection between two fluid reservoirs. The device is plasma cleaned for five minutes before immediately adding buffered salt solution and desiccating to remove bubbles. These nanocapillaries are a simple, cheap alternative to nanopores made by ion beam milling methods with each nanocapillary costing less than $1 and taking less than a minute to produce.

Our laser pulling program produces a range of nanocapillary sizes of 29 ± 8 nm as estimated from measuring 35 nanocapillary diameters by scanning electron microscopy.

2.2 Synthesis of DNA origami

We designed a flat DNA origami shape approximately 60×50 nm2 wide and two helices in height and with a 10×10 nm2 nanopore in the centre. The structure was synthesised by mixing scaffold and staple strands in a 1:10 stoichiometric ratio in 1xTE buffer and 14mM MgCl2 before heating to 65°C and cooling over 2 days. AFM imaging showed a high yield of the correctly folded structure.

2.3 DNA origami translocation events

For nanocapillaries with a resistance of less than approximately 20 MΩ we observe clear translocations when DNA origami is added to the negatively biased reservoir in a concentration of 0.5 ng/ L. Figure 3 shows typical translocation events. A bias of 500 mV was applied and the current signals were filtered at 30 kHz and sampled at 200 kHz. A solution of 0.5xTBE, 5.5mM MgCl2, IM KCl was used for all recordings.

2.4 Trapping of DNA origami

In capillaries with resistance greater than approximately 20MQ, individual DNA origami structures can be trapped onto the end of the capillary. Figure 4 shows one such trapping event forming a hybrid DNA origami-nanocapillary nanopore.

3 CONCLUSION

DNA origami is a powerful technique for the construction of designer nanopores. Nanocapillaries can be used both to detect the translocation of DNA origami structures and to trap the origami to form a hybrid pore by tuning the diameter of the nanopore. The detection of translocations is a potential new tool for assessing the folding quality and aggregation of DNA origami in solution. The formation of a hybrid origami-nanocapillary nanopore represents a novel method for creating hybrid nanopores with high throughput and low cost.

We thank Lorenz Steinbock for SEM imaging of nanocapillaries.

ON THE DEVELOPMENT OF NEW METHODS FOR ION CHANNEL STRUCTURE-FUNCTION MEASUREMENT

J.W.F. Robertson, V. Silin, and J.J. Kasianowicz

NIST, Physical Measurement Laboratory, Semiconductor and Dimensional Metrology Division, CMOS Reliability and Advanced Devices Group, Gaithersburg, MD 20899-1070

1 INTRODUCTION

One of the major roadblocks to understanding cell function, and the development of therapeutic agents against disease, is the relative paucity of tools for determining integral membrane protein (IMP) structures. Despite the extensive efforts applied to the problem, there are only ca. 300 solved structures of IMPs. Unfortunately, of those with solved structures, many are the same protein from different species. Considering the medical and technological importance of IMPs, it is clear that new technologies need to be developed to address this metrology issue.

There are four major techniques for obtaining high resolution structures of proteins: x-ray crystallography, electron microscopy, NMR spectroscopy, and EPR spectroscopy. Each of these measurement modalities has its own advantages and limitations. For example, when they can be obtained, x-ray crystal structures provide much greater detail (atom level) well beyond the resolution of electron microscopy. However, the method does not verify whether the structure is that of a fully functional molecule. This is a significant concern, because the environment required for stable functional IMPs (a lipid bilayer membrane) is a stark contrast to that required to isolate the protein from the membrane and then to coax it to crystallize. NMR can provide estimates of molecular motion at atomic resolution, but the method is currently limited to the study of relatively small proteins (< 32 kg/mol) or protein fragments confined to small volumes or on surfaces. The electron spin analog of NMR, EPR is used to elucidate membrane protein topology (e.g., the local structural motif of the protein: α-helix or β-sheet) and dynamics (e.g., whether a spin-containing amino acid side chain is in the membrane or aqueous phase).

While the contributions of these techniques to our present understanding of cell biology cannot be overstated, new technologies need to be developed that address corollary information and that can independently provide structural information and simultaneously provide evidence that the proteins actually work (there is essentially no value in knowing the structure of a non-functioning protein). Here, we discuss several such techniques, which are based upon the fabrication of a biomimetic membrane on a smooth metal electrode. The interface, a tethered bilayer lipid membrane (tBLM), is based upon self- assembly of lipid-like compounds to the electrode through thiol-chemistry with a spacer segment of a polymer. More recent approaches have used poly(ethylene glycol) as the spacer. Figure 1 (top) illustrates the membrane system in our experiments, which enables the use of a suite of surface-sensitive measurement modalities, including, but not limited to, electrochemistry (to test the protein's function) and a host of spectroscopic techniques to probe the structure of the molecule. A central goal of this system is to tailor the surface coverage of a given protein for the analytical method of choice.

2 METHODS AND RESULTS

For the study of bacterial pore forming toxins, one approach for optimizing the surface coverage of proteins is the simultaneous measurement of the resonance angle via surface plasmon resonance (SPR) and the transmembrane resistance via electrochemical impedance spectroscopy (EIS) as depicted in Fig. 1 (bottom). The advantage of combining these measurements is that they sample different aspects of the protein's behavior. Surface plasmons can exist at a metal surface, are highly confined near the interface, and are highly sensitive for the dielectric constant of a thin film at the metal surface. The SPR signal is estimated from the minimum of reflected light at the angle of incidence when x component of the incident light wave vector that of surface plasmons. When the optical properties of the interface change (e.g., due to protein adsorption) the resonance angle shifts. With appropriate assumptions about the optical constants of the surface and proteins, a measurement of these shifts is used to estimate the protein surface concentration. Conversely, EIS monitors changes in the ionic conductivity of the interfaces by applying a small AC potential wave and recording the current and phase shift over a wide range of frequencies, typically ~50 kHz to ~10 mHz. The data are subsequently analyzed by fitting to an idealized equivalent circuit model made of simple resistors and capacitors. For pore-forming toxins and other ion channels, the technique is highly sensitive with detection limits approaching the single molecule level, and provides a wide dynamic range to observe channel formation over many orders of magnitude of surface coverage (Figure 1 (top)).

Recently published work on channel formation by the pore-forming toxin Staphylococcus aureus α-hemolysin demonstrated the power of combining EIS and other measurement modalities (e.g., neutron reflectometry). Preliminary results for channel formation by the 63 kg/mol fragment (PA63) of protective antigen secreted by Bacillus anthracis provide a good example of the synergy of simultaneous SPR and EIS measurements. A typical SPR measurement shows that upon addition of the PA63, the SPR signal suggests that most of the protein adsorbs relatively quickly at the interface, but a slower adsorption process occurs as well (Fig. 2). The corresponding EIS signal (Fig. 3) initially shows a high membrane resistance (red) upon injection of protein into the bulk solution this resistivity drops smoothly as a function of time (blue, green, purple and black) with a corresponding shift in the frequency of the minimum phase shift. Taken together, these results demonstrate that some of the protein adsorbed to the membrane surface penetrates the membrane and forms ionic channels. However, a comparison of the mass of adsorbed protein (estimated from the SPR signal) to the surface coverage of ion channels (estimated from EIS and the known conductance of single PA63 ion channels, suggests that much of the adsorbed protein does not form channels. Although the SPR data indicate that the protein is absorbed at significant surface coverage (Fig.2), only a small fraction insert their P-barrel into the membrane are anchored to the surface, suggesting that part of the protein can denature and lose the ability to form channels. The added advantage here is that ion channels can be tested for function prior to, or simultaneously to, application of structural techniques (see below), which has not been demonstrated with any other technology.

Our initial neutron reflectometry (NR) experiments (data not shown) were hampered by low surface coverage of PA63 channels, but strongly suggest that this approach will ultimately succeed as it did for the α-hemolysin channel. As with x-ray reflectometry, the neutron intensity decreases dramatically as a function of momentum transfer. Data were collected for both the membrane and the protein-infused membrane under three different solvent contrasts -- neutrons interact with the nucleus of atoms and protons and deuterium offer the largest difference between two isotopes of any element that occurs in biological systems. Features in the data are dominated by the interference pattern from neutrons reflected off the various interfaces in the system. To develop a picture of the interface, layer-by-layer models must be constructed and tested against these data. These preliminary results suggest that the surface coverage is currently only between 1 % and 3 %, which does not provide sufficient detail to make significant claims about the structure of this pore-forming toxin.

3 CONCLUSION

Biomimetic interfaces can be constructed on atomically smooth electrode surfaces to provide a measurement platform that allows both high-information structural experiments such as NR, or vibrational spectroscopy as well as allow functional measurements of the protein as well. The true advantage of the robust tethered bilayer is the ability to make many different measurements over a long period of time on the exact same interface. No other technology affords this ability.

4 ACKNOWLEDGMENTS AND NOTES

Supported in part by the NIST Office of Law Enforcement Standards. Certain methods are mentioned in the manuscript. This in no way represents an endorsement by NIST.