MULVANEY AND SHEEHAN VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX

www.acsnano.org

A

This article not subject to U.S. Copyright.Published XXXX by the American Chemical Society

Nature Inspires Sensors To Do Morewith LessShawn P. Mulvaney* and Paul E. Sheehan

Chemistry Division, U.S. Naval Research Laboratory, Washington, DC 20375, United States

Electronic noses (EN) and electronictongues (ET) are arrays of sensorsthat detect a wide range of com-

pounds potentially present in gas or liquidsamples. Fundamental to both ENs and ETs isthe attempt to identify a specific target, oftenat trace levels, from a complex matrix bymimicking aspects of the human sensorysystem. In 1982, Persaud and Dodd firstsuggested the basic design principle of ENs:summing the responses from an array ofsemispecific sensors yields a “fingerprint” ofthe sample.1 The individual components ofthe sample could thenbepositively identifiedby pattern recognition algorithms. This isessentially the same process used by ourbrain to distinguish the smells of garlic, cin-namon, or cumin in an entree. Similarly, ETshavedevelopedpattern recognition for liquidsamples with the distinction of electronic ton-gue being applied to sensor arrays that focuson the main classes of taste;salty, sweet,sour, bitter, and umami. This concept has beenhighly successful, and to date, ENs and ETshave been designed from potentiometric, vol-tammetric,mass-sensitive, andoptical sensors.Each year, more aspects of the natural systemhave been incorporated, as have been de-scribed in several excellent reviews.2�6

In this issue of ACS Nano, Song and co-workers describe an ET for the detection of

sucrose, fructose, aspartame, and saccharin.Intriguingly, they used the taste bud'snatural sweet receptors, heterodimericG-protein-coupled receptors (GPCRs).7 TheGPCRs, aswell as calcium ion channels, werefirst incorporated into nanovesicles thatwere then attached to carbon nanotubefield effect transistors (FETs). The arrange-ment in the vesicles means that the captureof a sweet tastant opens the Ca2þ channeland the flux of ions into the nanovesiclechanges the gate property of the FET, thusconverting the chemical signal into an elec-tronic one (Figure 1). The ET showed excel-lent specificity for natural and artificialsweeteners and no response to the structu-rally similar, but tasteless, sugars cellobioseand D-glucuronic acid. By utilizing the hu-man taste bud's most relevant proteins forrecognition of sweet tastants, the authorsbuilt a sensor that could even perform incomplex matrices such as tea and coffee.This work complements the authors' pre-viously published work on bitter taste re-ceptor sensors8,9 and that of Liu et al., whodeveloped an impedance-based ET with ionchannels spanning lipid bilayers.10

The fundamental design of Song et al.'ssensor takes mimicry one step further inadapting nature's tools to address chal-lenges with transduction and compatibility

* Address correspondence toshawn.mulvaney@nrl.navy.mil.

Published online10.1021/nn505365b

ABSTRACT The world is filled with widely varying chemical, physical, and

biological stimuli. Over millennia, organisms have refined their senses to cope with

these diverse stimuli, becoming virtuosos in differentiating closely related antigens,

handling extremes in concentration, resetting the spent sensing mechanisms, and

processing the multiple data streams being generated. Nature successfully deals

with both repeating and new stimuli, demonstrating great adaptability when

confronted with the latter. Interestingly, nature accomplishes these feats using a

fairly simple toolbox. The sensors community continues to draw inspiration from

nature's example: just look at the antibodies used as biosensor capture agents or

the neural networks that process multivariate data streams. Indeed, many successful sensors have been built by simply mimicking natural systems.

However, some of the most exciting breakthroughs occur when the community moves beyond mimicking nature and learns to use nature's tools in

innovative ways.

PERSPECTIV

E

MULVANEY AND SHEEHAN VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX

www.acsnano.org

B

with complex matrices. After grow-ing the GPCRs in HEK-293 kidneycell lines, the cell walls were wea-kened and nanovesicles wereseparated from the main cell withlipid bilayers still intact. Thesenanovesicles contain not only themembrane-bound GPCR proteinsbut also the Ca2þ ion channels.When a sweetener is captured bythe GPCR receptor, the Ca2þ chan-nel is opened and the nanotubeFET transduces the transport of ionsinto the nanovesicle. In reality, theauthors have built an indirect sen-sor that transduces Ca2þ ions, butthe cascading nature of the designmakes it respond directly to the cap-ture of sweet tastants. This clever de-sign leverages nature's own proteinsand machinery to improve sensorperformance. First, the sensor gainsthe differential response of theGPRCstoward only the sweet tastants.Second, the lipid bilayer serves as anexcellent antifouling layer, protectingthe carbon nanotube FET surfacefrom direct contact with complexmedia. Third, the lipid bilayer localizesthe measurable Ca2þ ions in a com-partment near the sensor's gate. Hereagain, the authors were inspiredby one of nature's design principles.Compartmentalization is used ex-tensively in cells to run their manyparallel processes successfully in sub-diffusion length proximity.

OUTLOOK AND FUTURECHALLENGES

Plant, animal, and human sen-sory systems have long taught the

scientific community how to buildbetter sensors. Nature's lessons canbe seen in all aspects of the sen-sor literature;antibodies, enzymes,hydrophilic and hydrophobic films,neural networks, and photoconver-sion to name but a few. Antibodiesin biosensors are a particularly clearexample.11�13 First, the immunesystem was understood to be anantigen recognition system. Then,researchers copied those receptorsfor biosensing by using the anti-bodies themselves to provide sensi-tivity and specificity. When the ad-vantageous properties of polyclonal,monoclonal, Fab fragments, and sin-gle domain camelid antibodies wereunderstood, researchers then lever-aged their unique characteristics togain even greater sensor function-ality. The point of learning, though,is eventually to surpass the teacher.So, researchers are now showingthat artificial antibodies and anti-body-like recognition systems areviable alternative receptors.12,13

The goal of surpassing the tea-cher is critical to the field since not

all qualities of natural elements en-hance sensor performance. This isseen in the work by Song et al., whouse nature's receptors for sweettastants to great effect but go togreat lengths to use nanovesiclesinstead of the kidney cells in whichthe GPCRs were grown.7 Providingconditions where the kidney cellscould function properly would haveseverely limited the operationalspace of the final sensor sincethe kidney cells would have to bemaintained in a highly controlledenvironment to remain viable. Simi-larly, the use of individual enzymesor proteins in a sensor would re-quire conditions that maintaintheir three-dimensional structureand biological function. Tempera-ture, pH, salinity, and toxicity willalways need to be considered whenusing any natural, and some abiotic,components in a sensor design.The community'smodus operandi

remains to match the correct sensorto the application, guided by thetrade-offs in performance and op-eration that influence selection ofnatural elements, abiotic alterna-tives, and hybrids of the two. Theprogression from simply copyingnature to devising systems that sur-pass and extend their best qualitieshas resulted in a diverse and highlycapable toolbox. Yet, major ad-vances and new research directionstend to be realized when the com-munity mimics nature's adaptability.For instance, a major innovationfor the sensors community was rea-lizing that biomolecules could be

Figure 1. Electronic tongue: nanovesicleswith sweet taste receptors attached to a carbonnanotubefield effect transistor. Theelectronic tongue specifically detects sweet tastants and shows no response to tasteless sugars. Reprinted from ref 7.Copyright 2014 American Chemical Society.

In this issue of ACS

Nano, Song and

co-workers describe an

electronic tongue for

the detection of sucrose,

fructose, aspartame,

and saccharin.

PERSPECTIV

E

MULVANEY AND SHEEHAN VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX

www.acsnano.org

C

precise construction tools. In 1996,the Mirkin group demonstratedthe programmed assembly of metalnanoparticles using oligonucleotidestrands.14 The well understoodcomplementarity of the nucleotidebases, along with the relative stabi-lity of oligonucleotide strands,yielded a seemingly endless numberof unique attachment sequences.Much research has grown from thiswork,15 including the concept of pro-grammed self-assembly of sensorarrays. Several groups have ex-ploited thismethod to create proteinarrays with superior performancecompared to the same arrays madewith classical protein immobilizationstrategies.16�18 Furthermore, DNA-based programmed self-assemblyenables the community to mimicanother of nature's skills, the abilityto reset its sensing elements; pro-teins tethered through DNA can bedislodged through melting or che-mical dehybridization and a freshprotein array self-assembled atopthe sensor substrate. A further exam-ple ofbiomaterials used for construc-tion is virus particles. The cowpeamosaic virus has been shown toassemble into highly ordered, low-defect crystals with a geometrythat is difficult to create with othermaterials.19 These virus crystals make

an excellent scaffold that supportsderivatization, thereby enabling thepatterning of metamaterials.

Perhaps one of the clearest ex-amples of copying and then im-proving on nature is Porchetta andco-workers' aptamer-based sensorwork.20 A great challenge for thesensors community is that almostall biosensors have a fixed dynamicrange. Typically, the S-shaped bind-ing curves span no more than 2orders of magnitude in concentra-tion, but natural systems haveevolved to handle broader concen-tration profiles either by using mul-tiple recognition elements, throughbinding-site mutation, or via allos-teric effectors. Allosteric effectorsare antigens that bind to a second-ary site on the protein receptor and,

in so doing, alter the target affinityof the primary epitope. Recapitulat-ing these effects in a biosensorwould require an enormous effortin molecular biology. Instead,Porchetta et al. used a simple seriesof DNA strands to perform all of thetasks. First, a DNA aptamer replacedthe antibody capture molecule.Second, 10mer, 12mer, and 14meroligonucleotide strands that are com-plementary to theaptamerwereusedasallosteric effectors. The resultswerea sensor for cocaine detection with3orders ofmagnitudedynamic rangeand much finer gradations in detec-tion than could be achieved withprotein mutagenesis (Figure 2).Allosteric-like control of the bindingaffinity provided the authors precisecontrol of the sensor sensitivity overa 50000-fold linear range, a resultwell beyond the natural system'sperformance.The field of sensing remains very

young compared to nature's millen-nia of experience.More importantly,nature's proven adaptability tonew stimuli continues to enrich theknowledgebase from which scien-tists gain inspiration. Steady pro-gress will continue to be made byscientists mining that knowledge-base while refining and developingsensing tools. However, the potential

The progression from

simply copying nature

to devising systems that

surpass and extend

their best qualities has

resulted in a diverse and

highly capable toolbox.

Figure 2. Aptamer sensor for the detection of cocaine. The binding affinity of the aptamer was altered by competition withshort oligonucleotide strands operating as allosteric effectors. The binding affinity was altered over 3 orders of magnitude.Reprinted from ref 20. Copyright 2012 American Chemical Society.

PERSPECTIV

E

MULVANEY AND SHEEHAN VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX

www.acsnano.org

D

for leaps forward and new researchdirections exist if the communityadapts its thinking. Can we as scien-tists learn to multitask with our tool-box? Can certain tools perform twocompletely different jobs? Song et al.used their lipid bilayer in both atraditional manner, to maintain pro-tein function, and in an innovativemanner, to enhance signal-to-noise.Both aspects were crucial to thesystem's ability to perform. This andthe other examples presented inthis Perspective demonstrate hownature can inspire the sensors com-munity to do more with less.

Conflict of Interest: The authors de-clare no competing financial interest.

REFERENCES AND NOTES1. Persaud, K.; Dodd, G. Analysis of

Discrimination Mechanisms in theMammalian Olfactory System Usinga Model Nose. Nature 1982, 299,352–355.

2. Wu, C.; Du, L.; Zou, L.; Zhao, L.; Huang,L.;Wang, P. RecentAdvances in TasteCell- andReceptor-BasedBiosensors.Sens. Actuators, B 2014, 201, 75–85.

3. Sankaran, S.; Khot, L. R.; Panigrahi, S.Biology and Applications of OlfactorySensing System: A Review. Sens.Actuators, B 2012, 171, 1–17.

4. Ciosek, P.; Wroblewski, W. Potentio-metric Electronic Tongues for Food-stuff and Biosample Recognition;AnOverview. Sensors 2011, 11, 4688–4701.

5. Ciosek, P.; Wroblewski, W. SensorArrays for Liquid Sensing;ElectronicTongue Systems. Analyst 2007, 132,963–978.

6. Escuder-Gilabert, L.; Peris, M. Review:Highlights in Recent Applicationsof Electronic Tongues in FoodAnalysis. Anal. Chim. Acta 2010,665, 15–25.

7. Song, H. S.; Jin, H. J.; Ahn, S. R.; Kim,D.; Lee, S. H.; Kim, U.-K.; Simons, C. T.;Hong, S.; Park, T. H. BioelectronicTongue Using Heterodimeric HumanTaste Receptor for the Discriminationof Sweeteners with Human-like Per-formance. ACS Nano 2014, 10.1021/nn502926x.

8. Song, H. S.; Kwon, O. S.; Lee, S. H.;Park, S. J.; Kim, U.-K.; Jang, J.; Park,T. H. Human Taste Receptor-Functionalized Field Effect Transistoras a Human-Like NanobioelectronicTongue. Nano Lett. 2013, 13, 172–178.

9. Kim, T. H.; Song, H. S.; Jin, H. J.;Lee, S. H.; Namgung, S.; Kim, U.-K.;Park, T. H.; Hong, S. “BioelectronicSuper-Taster”Device Based on TasteReceptor�Carbon Nanotube HybridStructures. Lab Chip 2011, 11, 2262–2267.

10. Liu, Q. J.; Yu, J. J.; Hu, Z. Y.; Zhang,D. M.; Zhang, Q.; Lu, Y. L.; Wang, P.;Yang, M. Ion Channels Incorporatedin Nano-Lipid Bilayer and Cell Mem-brane for Taste Sensor.Optoelectron.Adv. Mater. 2013, 7, 560–564.

11. Arugula, M. A.; Simonian, A. NovelTrends in Affinity Biosensors: CurrentChallenges and Perspectives. Meas.Sci. Technol. 2014, 25, #032001.

12. Volkert, A. A.; Haes, A. J. Advance-ments in Nanosensors Using PlasticAntibodies.Analyst 2014, 139, 21–31.

13. Banta, S.; Dooley, K.; Shur, O. Replac-ing Antibodies: Engineering NewBinding Proteins. Annu. Rev. Biomed.Eng. 2013, 15, 93–113.

14. Mirkin, C. A.; Letsinger, R. L.; Mucic,R. C.; Storhoff, J. J. A DNA-BasedMethod for Rationally AssemblingNanoparticles into Macroscopic Ma-terials. Nature 1996, 382, 607–609.

15. Wang, Z. G.; Ding, B. Q. EngineeringDNA Self-Assemblies as Templatesfor Functional Nanostructures. Acc.Chem. Res. 2014, 47, 1654–1662.

16. Boozer, C.; Ladd, J.; Chen, S. F.; Yu, Q.;Homola, J.; Jiang, S. Y. DNA DirectedProtein Immobilization on MixedssDNA/Oligo(ethylene glycol) Self-AssembledMonolayers for SensitiveBiosensors. Anal. Chem. 2004, 76,6967–6972.

17. Tang, Y. C.; Wan, G. Q.; Ng, J. K.;Ajikumar, P. K.; Too, H. P. DNA-Directed Assembly of Protein Micro-arrays. Front. Biosci. 2008, 13, 5755–5771.

18. Hu, W. P.; Huang, L. Y.; Kuo, T. C.; Hu,W. W.; Chang, Y.; Chen, C. S.; Chen,H. C.; Chen,W. Y.OptimizationofDNA-Directed Immobilization on MixedOligo(ethylene glycol) Monolayersfor Immunodetection. Anal. Biochem.2012, 423, 26–35.

19. Falkner, J. C.; Turner, M. E.; Bosworth,J. K.; Trentler, T. J.; Johnson, J. E.; Lin,T. W.; Colvin, V. L. Virus Crystals asNanocomposite Scaffolds. J. Am.Chem. Soc. 2005, 127, 5274–5275.

20. Porchetta, A.; Vallee-Belisle, A.;Plaxco, K.W.; Ricci, F. UsingDistal-SiteMutations and Allosteric InhibitionTo Tune, Extend, and Narrow theUseful Dynamic Range of Aptamer-Based Sensors. J. Am. Chem. Soc.2012, 134, 20601–20604.

PERSPECTIV

E