My interest in small-scale single-molecule biology emerged from consulting while doing my post-graduate research. I found the ability of deliberately designed nanostructures to modulate and transduce single molecules a powerful and elegant phenomenon to explore. To that end, I did my MSEE work in the Nanopore Physics Lab at the University of Washington.
Nanopores are, in caricature, a nanoscale extension of the well-understood Coulter counter. A single hydrated channel, nanometers in diameter, acts as a passage between two volumes of conductive solution. This channel modulates and detects the translocation of single molecules under appropriate electrical bias by virtue of the sample molecules perturbing the ion current through the channel. The most well-understood channel-forming technique is the insertion of a single protein porin in a suspended lipid bilayer, with a research history marked by the 1991 Nobel Prize for the development of the patch-clamp amplifier. An older, but particularly well-written review can be found here.
The most pressing application of nanopores the development of nanopore-based DNA sequencing tools. A single-stranded DNA molecule translocating though an appropriate pore can modulate the translocation current as each base moves though the porin constriction zone; these current signatures can be used to decode the sequence.
To form nanopores, the first commonly used protein was alpha hemolysin, a channel-forming protein from Staphylococcus aureus. This protein consists of an exterior (trans-side) lumen leading to a short beta-barrel; the tight confinement over substantial length occludes translocation current signatures. On the other hand, the MspA protein from Mycobacteria smegmatis contains a much narrower constriction zone, which in turn yields much higher current resolution for a single-strand DNA translocation experiment.
While contemporary protein engineering techniques engender substantial flexibility in modifying the channel geometry and charge location, this architecture suffers from poor mechanical and electrical robustness. A serious competitor has emerged – the solid state nanopore. This nanopore is a single 2-3 nanometer aperture ablated in a solid-state supporting membrane. While not as finely tuned to specific DNA-protein interactions, solid state nanopores are much more mechanically robust, and a suite of approaches exists to ameliorate the poor sensitivity. Moreover, membrane materials have evolved from silicon nitride to other materials – aluminum nitride and graphene, as well as multilayered systems, stand out as particularly interesting examples – enabling more elaborate solutions to the issues of sensitivity and facility of use.
Below, I outline the two papers I published on the use of a novel bilayer chemistry as a more robust alternative, and the initial steps I took to hybridize MspA and a solid-state nanopore, potentially marrying the strengths of the two systems.
Protein nanopores are supported by lipid bilayers; in living systems, these self-assemble from free lipids to form cell membranes. In a nanopore experiment, the lipid bilayer is spread over an aperture in a PTFE membrane or tube to form a substrate for porin insertion. Since nanopores operate under electrical bias, the breakdown voltage – the maximal transmembrane voltage sustainable by a lipid bilayer – is important to the measurement’s ultimate dynamic range. While typical lipids – phosphatidylcholines (PC) – on typical twenty-micron apertures break down at an acceptable 0.2 volts, the ability to drive at higher voltages enables a suite of novel experiments in using nanopores as force sensors.
To this end, I explored, co-developed and published on a novel bilayer chemistry with a breakdown voltage twice the previously set limit for porin-bearing chemistries. Inspired by the natural environment of the MspA protein, we explored mycolic acids from the cell wall of Mycobacteria smegmatis as a viable alternative to the PC lipids used in previous experiments.
Mycolic acid is a unique molecule in that the carboxylic acid head group is well-defined, but the two aliphatic chains are variable, yet long and asymmetrical. Each molecule contains 60 to 90 carbon atoms, but the relative chain length varies by a factor of two. This leads to a unique – and heretofore not fully understood – packing arrangement, which in turn engenders a particularly robust cell membrane. This chemical and mechanical robustness was of particular interest to us in nanopore experiments using MspA, as stability and protein insertions were not as ready to come by as with hemolysin. Inspired by the native environment of the MspA in the M. smegmatis cell wall, we set out to reproduce this environment, and create the first synthetic mycolic acid bilayers.
Specifically, we developed the first unsupported bilayers in MyA and demonstrated gigaohm seals in ionic solutions. We developed a protocol for MspA porin insertion in the mycolic acid bilayers, and demonstrated that the protein structure and base-sensitivity upon translocation remained the same as with conventional lipids. This effort taught me about the realities of synthetic bilayer experiments, and of the capabilities of this platform. This work yielded two publications in 2010 and 2011, as well as a talk at the Biophysical Society Annual Meeting in 2010.
There exist two major implementations of the nanopore concept: the solid-state nanopore is robust and compatible with CMOS fabrication techniques, while the protein nanopore is more fragile yet provides finer control of the moieties interacting with the translocating polymers. It is a natural idea to seek systems that marry the strengths of the two approaches: one such possible tool may be a hybrid nanopore. In this system, a protein porin acts as a liner for a solid-state nanopore – the inner moieties of the protein interact with the translocating polymer and provide sensitivity, while the supporting outer membrane acts as a robust mechanical support.
This concept was demonstrated in 2010; to achieve similar results with MspA, we had to first achieve similar results in the implementation of solid-state nanopores, as the UW lab had not previously worked with this system. Thus, I developed a solid-state nanopore setup from scratch: I developed the nanopore sculpting process in silicon nitride membranes to reliably produce 2-3 nanometer pores. I developed a PDMS and teflon chuck to carry the chip, and reliably insulate the two sides electrically and from ionic current. I developed a protocol to wet the nanopores – silicon nitride is hydrophilic. I developed a protocol for translocating double-stranded DNA though the nanopore, and demonstrated the first double-stranded DNA translocating though a nanopore in the UW lab. I then experimented with functionalizing the silicon nitride nanopores, to anticipate solutions to problems with sealing the MspA protein in the silicon nitride aperture.
With the development of the silicon nitride system complete, I turned to the biochemistry required to insert the protein into the aperture. We performed experiments with bare MspA, and saw no reliable controlled insertion evidenced by a conductance falling to the accepted MspA value. This is expected, since the MspA carried no substantial dipole moment, and should not orient successfully in the applied electric field across the nanopore. Instead, we hypothesized that streamers, consisting of nucleic acids, attached to the bottom of the MspA molecule would guide the protein in, and orient it correctly to the silicon nitride aperture.
To do these experiments, we performed a point substitution of the amino acids building up the protein, introducing a cysteine in the cis-side of the porin. Since MspA has no naturally occurring cysteines, this point substitution would provide a target for a thiol-thiol disulfide bridge. I was able to developed a disulfide bridge protocol to link thiolized DNA and the resultant cysteine residue. This approach suffered from a substantial failure mode: MspA is expressed as an octomer – a mutation in a given subunit is expressed eight times, leading to eight possible attachment sites for the thiolized DNA. This is problematic, but easily solved by brute force – performing the ligation procedure, then using gel electrophoresis to separate the products with an overabundance of ligated DNA. To slightly increase yield, we also added only sufficient DNA to complete the ligation on only one of the target sites per porin.
With this protocol in place, we proceeded to look at various streamers to attach to the protein. In our insertion experiments, we considered single-stranded and double-stranded DNA. We used Watson-Crick basepairing to bring together longer strands, from the phage-lambda genome and ligate them to the single-stranded molecule tied to the bottom of MspA.
These insertion experiments yielded promising early results that are explained in further detail in my thesis. Before the conclusive completion of the work, I was recruited by a nanoscale optics startup, and went on leave from the UW.