Nanodevices for advances in single molecule analysis

S. Kirkwood, 2018

Nanobioelectronics explores the interface between biological and technological systems. Nature
possesses a great many versatile molecules and processes, the detailed understanding of which
would surely lead to huge advancements in medicine, computing, biotechnology and numerous other
fields. This review discusses the use of nanodevices including nanowire and nanotube field effect
transistors in analysis and sensing of biological systems, and the significance of single molecule
analysis in studying the complexities of biological environments.

Introduction
Millions of years of evolutionary streamlining has provided nature with an impressive arsenal of
biological instruments. From the ubiquitous ATP synthase, an enzyme which is ‘one of the most
beautiful as well as one of the most unusual and important’1, to motor proteins, molecular machines
which perform a vast number of essential tasks in cells including replication and intracellular
transport2, the variety and efficiency of biological systems is astonishing. Many of the characteristics
of biological systems are ones that manmade systems aim to emulate – speed, efficiency, specificity
and small size. Yet, despite the rapid pace of technological development over the last few decades,
humanity’s most sophisticated machines are far behind the elegant solutions found in nature 3. This
disparity is shown in the endeavour to simulate the human brain using a supercomputer. In 2018, a
team of researchers based in Germany, Japan, Norway and Sweden developed software that
theoretically should be able to simulate the brain’s 100 billion neuron connections4. However, at
present there is no supercomputer capable of running the simulation5. In 2013, a joint project
between Japanese and German researchers used RIKEN’s K supercomputer, the fourth most powerful
supercomputer at the time (currently sixteenth in the world6), to run a simulation of just 1% of the
brain’s neuronal activity. The simulation utilized 700,000 processor cores and 1.4 million gigabytes of
RAM, and yet it took 40 minutes to model just one second of brain activity7. Furthermore, the
computer comprises some 800 cabinets of circuitry, and consumes 9.89MW – or the equivalent of
around 10,000 homes. Clearly, there are many challenges in creating technology that can rival that of
biological systems.

In response to this challenge, many groups of researchers internationally have turned their attention
to combining biological and electronic systems. Nanobioelectronic devices could potentially have
significant impacts in biotechnology, medicine and computing8. Already, numerous devices have been
fabricated that have allowed enhanced measurement of biological molecules – for example Misra et
al. developed a silicon nanowire device incorporating a lipid bilayer membrane and functional
membrane proteins which is able to convert biological changes into electrical signals8, and Sorgenfrei
et al. were able to examine the kinetics of DNA hybridisation by attaching a probe DNA molecule to a
point defect created on a carbon nanotube9. However, this rapidly expanding field comes with its own
obstacles – largely, the difficulties of interfacing biological molecules with electronic circuitry3.

Biological systems operate under a vastly different paradigm to modern electronics – where circuits
operate via electron currents, natural systems use flows of small molecules, ion gradients and
membrane electric potentials to send signals8. Furthermore, knowledge of the precise functioning of
many key biological elements is still fuzzy, with single molecule analysis still an emerging field
providing much needed clarity over past studies which have sampled aggregate properties of large
numbers of molecules at a time9,10. Noy asserts in his review of nanobioelectronics that there are three
essential properties for successfully integrating biological and electrical systems – first, there must be
good coupling between the organic and inorganic components in a biological environment; second,
the device must have good signal to noise ratio; and third, the technology must be simple and
versatile. This review will discuss the efforts of many research groups in creating nanobioelectronic
devices, and their potential applications in a variety of fields including single molecule analysis.

Nanodevices
One of the biggest challenges in interfacing electronics to biological systems is solving the size
mismatch between biologically important molecules and typical electronic components12. The usual
electronic toolkit of macroscopic wires, sensors and circuit components is simply too large to be
applied to the nanoscale world of proteins, membranes and DNA. Nanoscale materials developed over
the last few decades provide the solution. Nanotubes and nanowires are the ideal mix between nano
and macro – down to a fraction of a nanometre in width yet long enough to be seen with the naked
eye in some cases. This quasi one-dimensionality makes them uniquely suited to nanobioelectronic
applications due to several key properties. Firstly, their high surface to volume ratio allows for
excellent sensitivity, so that only a small number of analyte molecules are required to produce a
measurable signal12. Secondly, they are of a size scale comparable to that of biological building blocks3,
as shown in figure 1 below. Additionally, unlike materials which are nanoscale in all directions
(quantum dots), their macroscopic length allows them to be interfaced with circuitry components in
order to detect the electric signals11.

Figure 1: comparison
of the sizes of
essential biological
molecules and
various
nanomaterials.
Image from Noy,
Bionanoelectronics3.

1. Nanowires
Nanowires are solid, crystalline fibres with a diameter typically between 3 – 500nm and a length
ranging from nanometres to millimetres13. Nanowires can be fabricated from a variety of materials –
metals, semiconductors, insulators and organic compounds11, though are typically made from silicon.
This is due to the silicon industry being so well established, meaning that properties such as
morphology, size and doping can be readily controlled12. Also, the silicon oxide surface of silicon
nanowires is very hospitable to biological molecules3, a very useful feature for biological applications.

2. Nanotubes
Nanotubes are rolled up graphene sheets, and can be classified as either single-walled or multi-walled.
Their high mechanical strength and chemical and thermal stability provide them with excellent
electronic properties12. Furthermore, the diameter of a nanotube is typically 0.4 – 2nm, smaller than
that of a nanowire. However, their impressive qualities come at a cost – nanotubes are somewhat
more difficult to work with than nanowires. Their electronic properties are determined by their
chirality, that is, the angle at which the sheets roll up with respect to lattice vectors14. Depending on
the chirality, nanotubes can either behave as a metal or as a semiconductor. With current fabrication
techniques, approximately one third of fabricated nanotubes are metallic and the other two thirds are
semiconducting15. For nanoelectronic applications, semiconducting nanotubes are needed12, and thus
additional steps of separating the two varieties are required.

3. Field Effect Transistors (FET)

In addition to being the most fundamental unit of all modern computers, the versatile transistor also
has applications in biosensing. Figure 2 below shows the structure of a typical field effect transistor.
The flow of charge carriers in the semiconducting material between the source and the drain is
modulated by a voltage applied to the gate.

Figure 2: cross section of a field

effect transistor, showing the
semiconducting channel and the
source, drain and gate
electrodes. Image from Serges16.

For biosensing applications, the channel between the source and drain electrodes is made from one
or multiple nanowires or nanotubes. Rather than a gate electrode, nanowire or nanotube FET devices
translate events in the vicinity of the transistor to electrical signals through charge gating – that is,
charged molecules or processes that alter the electrical properties of the nanowire/tube modulate
the current flow through the transistor in a similar fashion to an applied voltage at a normal gate
electrode3. Such devices fulfil the requirements listed above for successful integration into a biological
environment – they are able to operate in ionic solutions, the technology is relatively simple and very
versatile with many potential applications, and the devices are also able to achieve good signal to
noise ratio, though a limitation is the need to find a compromise between conductance, sensitivity
and ease of fabrication. FET devices can be made more sensitive by reducing the amount of doping
however, thin nanowires with low doping are difficult to make and also hard to pass current through3.
Despite this, many groups have fabricated devices which are able to be applied to a range of
biosensing tasks.

In order to better interface with the biological world, FET devices often utilise lipid bilayer membranes
as a fundamental component. Lipid bilayer membranes are found in virtually all cells on Earth, as they
form a selectively permeable barrier between the intracellular and extracellular environments. They
provide a natural structure in which to incorporate functional membrane proteins, which in cells are
responsible for detecting environmental changes and transporting materials in and out of the cell.
Misra et al. created a device which uses shielded nanowires coated in a lipid bilayer (figure 3) in order
to achieve ionic to electronic signal transduction. The group demonstrated incorporating two different
functional membrane proteins into the bilayer to introduce sensing capabilities to the FET device8.
First, gramicidin A, a helical polypeptide which allows the passage of monovalent cations through a
membrane, was incorporated into the bilayer. It was shown that the gramicidin A molecules were able
to let hydrogen ions pass through the membrane, causing a pH change at the surface of the nanowire
which in turn triggered a change in conductance. This effect was sharply diminished upon the addition
 of calcium ions, which bind near the mouth of the gramicidin A molecules and inhibit the transport of
hydrogen ions. A second device was made with alamethicin (ALM). ALM only forms pores through a
membrane when a voltage is applied. Misra et al. found that with no applied voltage, the device was
‘turned off’, and had no pH response. However, once a voltage was applied, the ALM pores allowed
hydrogen ions through in much the same way as gramicidin A.

B
A

Figure 3: A) diagram showing the structure of the shielded nanowire FET device, and the incorporation of the helical
gramicidin A across the membrane; B) conductance in response to pH change for a FET incorporating gramicidin A –
the red line shows the uncoated wire response, the blue line shows the response for a shielded wire with gramicidin
A, and the black line is the response once calcium ions are added. Image adapted from Misra et al8.
Many other groups have also had successes in creating shielded nanowire/tube devices. Huang et al.
demonstrated a carbon nanotube FET device which incorporates Na+/K+ ATP-ase, a membrane protein
powered by ATP hydrolysis to modulate output current by up to 40%17. This project is particularly
interesting in that the membrane protein used utilises active transport rather than passive transport
to translocate ions across the membrane. Hemmatian et al. also developed a shielded nanowire device
incorporating gramicidin A and ALM, similarly showing a chemically gated and voltage gated pH
response respectively18. Many of these groups comment that this type of device is a versatile platform
with many potential applications, including biosensing, diagnostics, neuroprosthetics and computing8.

Single-molecule analysis
Single-molecule analysis involves the study of individual molecules, rather than the aggregate
properties of many molecules. It is an expanding field which provides further insight into the action
of many fundamental biological molecules. For example, it was long thought that molecular motors
move continuously at a constant velocity, however single molecule analysis revealed that they in fact
take discrete ‘steps’10. Advances in these biological studies have come from the field of physics –
technologies such as optical tweezers and atomic force microscopy have been invaluable10.
Watanabe et al. developed a novel device for single molecule analysis by creating an arrayed lipid
bilayer chamber system19. The device comprises an array of 4µm diameter wells, each sealed with a
lipid bilayer. This setup allowed them to study the pumping action of the membrane proteins α-
hemolysin and ATP synthase. The studies involving ATP synthase were particularly significant, as it is
such an important molecule and single molecule analysis of its hydrogen ion pumping action had
never before been conducted. By introducing a solution containing molecules of ATP synthase,
Watanabe et al. were able to isolate one or two molecules per chamber and examine the change in
colour of a pH indicator inside the sealed chamber as the ATP synthase molecules translocated
hydrogen ions against the concentration gradient by ATP hydrolysis. The group found the proton
pumping rate of a single ATP synthase molecule to be 27.5 protons per second19.
 A Figure 4: A) structure of the
B
bilayer sealed chamber
containing a single ATP synthase
molecule The chamber is filled
with a pH indicator to allow
optical observation of ATP
synthase’s proton pumping
action; B) fluorescence images of
the wells containing ATP synthase
before and after the addition of
an ATP solution. Image adapted
from Watanabe et al19.

Future directions
The technologies discussed above present a broad platform for nanoscale analysis of biological
systems. Nanowire and nanotube devices clearly have many applications in responding to changes in
biological systems due to their size and electrical properties, however, in many of the studies
mentioned above, only the overall effect of many target molecules over a time interval was studied
(figure 3B). This limits the usefulness of the devices as a nuanced understanding of the membrane
proteins can only be obtained when studied at the single molecule level – as discussed above, the
efforts of Watanabe et al. in a novel analysis of single molecules of ATP synthase resulted in a much
more precise survey of its pumping action19. Thus, it seems that a desirable path of study would be
to combine these nanowire and nanotube devices with technologies that allow for analysis at the
single molecule level. One implementation of this would be to position a nanowire FET at the base of
each of the wells described by Watanabe et al. This would allow for simultaneous electronic and
optical detection of the action of membrane proteins, and would provide a detailed picture of the
function of the protein, with potentially far reaching, multidisciplinary impacts.

References
1. Boyer, P. (1997). THE ATP SYNTHASE—A SPLENDID MOLECULAR 10. Nature Methods. (2008). The smaller the better. [online] Available
MACHINE. Annual Review of Biochemistry, 66(1), pp.717-749. at: https://www.nature.com/articles/nmeth0608-457 [Accessed 31
2. van den Heuvel, M. and Dekker, C. (2007). Motor Proteins at Work for Aug. 2018].
Nanotechnology. Science, 317(5836), pp.333-336. 11. David L. Chandler, M. (2018). Explained: Nanowires and nanotubes.
3. Noy, A. (2010). Bionanoelectronics. Advanced Materials, 23(7), [online] MIT News. Available at: http://news.mit.edu/2013/explained-
pp.807-820. nanowires-and-nanotubes-0411 [Accessed 31 Aug. 2018].
4. Jordan, J. et al. (2018). Extremely Scalable Spiking Neuronal Network 12. Zhang, A. and Lieber, C. (2015). Nano-Bioelectronics. Chemical
Simulation Code: From Laptops to Exascale Computers. Frontiers in Reviews, 116(1), pp.215-257.
Neuroinformatics, 12. 13. Park, W. et al. (2008). Controlled Synthesis of Millimeter-Long
5. Tiwari, A. (2018). This Algorithm Can Simulate Our Brain At 100%, But Silicon Nanowires with Uniform Electronic Properties. Nano Letters,
No Supercomputer Can Run It. [online] Fossbytes. Available at: 8(9), pp.3004-3009.
https://fossbytes.com/brain-simulation-algorithm-100-percent/ 14. Ouyang, M. et al. (2003). Fundamental Electronic Properties and
[Accessed 31 Aug. 2018]. Applications of Single-Walled Carbon Nanotubes. ChemInform, 34(8).
6. Top500.org. (2018). TOP500 List - June 2018 | TOP500 Supercomputer 15. Saito, R. et al. (1992). Electronic structure of graphene tubules
Sites. [online] Available at: based onC60. Physical Review B, 46(3), pp.1804-1811.
https://www.top500.org/list/2018/06/?page=1 [Accessed 31 Aug. 16. Zambou, Serges. (2016). Electrical performance and use in logic of
2018]. current switching transistors employing nanostructured silicon.
7. Riken.jp. (2018). Largest neuronal network simulation achieved using 10.13140/RG.2.2.34537.90726.
K computer | RIKEN. [online] Available at: 17. Huang, S. et al. (2010). Carbon Nanotube Transistor Controlled by a
http://www.riken.jp/en/pr/press/2013/20130802_1/ [Accessed 31 Aug. Biological Ion Pump Gate. Nano Letters, 10(5), pp.1812-1816.
2018]. 18. Hemmatian, Z. et al. (2016). Electronic control of H+ current in a
8. Misra, N. et al. (2009). Bioelectronic silicon nanowire devices using bioprotonic device with Gramicidin A and Alamethicin. Nature
functional membrane proteins. Proceedings of the National Academy of Communications, 7, p.12981.
Sciences, 106(33), pp.13780-13784. 19. Watanabe, R. et al. (2014). Arrayed lipid bilayer chambers allow
9. Sorgenfrei, S. et al. (2011). Label-free single-molecule detection of single-molecule analysis of membrane transporter activity. Nature
DNA-hybridization kinetics with a carbon nanotube field-effect Communications, 5(1).
transistor. Nature Nanotechnology, 6(2), pp.126-132.