bioe 150 article 3

8/2/2019 BioE 150 Article 3

1/10


2/10

616 NATURE NANOTECHNOLOGY | VOL 6 | OCTOBER 2011 | www.nature.com/naturenanotechnology

the presence o di erent nucleotides are likely to be overwhelmedby thermodynamic uctuations (that is, statistical variations in thenumber o charge carriers and the position o the nucleotide in thepore). It has, there ore, proven impossible to sequence reely trans-locating ssDNA using -haemolysin.

Most nanopore sequencing strategies so ar have sought toactively or passively slow down the transport o ssDNA be orereadout. Active approaches typically incorporate enzymes to regu-late DNA transport through the pore1821. An enzyme motor coupledto a nanopore is attractive or two reasons: (1) the enzymeDNAcomplex orms in bulk solution, enabling it to be electrophoretically captured in the nanopore; (2) relatively slow and controlled motionis observed as the enzyme processively steps the DNA moleculethrough the nanopore. An elegant demonstration o this is thebase-by-base ratcheting o ssDNA through -haemolysin catalysedby DNA polymerase19. More recently, the processive replication o ssDNA on -haemolysin using phi29 DNA polymerase was demon-strated20. As well as being able to resolve individual catalytic cycles,polymerase dynamics (such as deoxyribonucleotide triphosphate (dN P) binding and the opening and closing o polymerase ngers)

could also be discerned rom measurements o the ionic curreWork on the controlled transport o DNA through -haemolysusing enzymes was reviewed recently 13.

Simpler, passive approaches to slowing down DNA also exissuch as nucleotide labelling, end termination o ssDNA with Dhairpins and the use o positively charged residues in the nanopas molecular brakes22 but all require urther work. Nucleotidlabelling, or example, is attractive, as the possibility o varyingchemistry, charge and size o the label might enable direct, real-tsequencing23, but it remains challenging to label contiguous nucletides in large genomic ragments.

A more versatile, label- ree sequencing method was recendemonstrated using resistive current measurements to cotinuously resolve indigenous mononucleotides (2 -deoxyadenosine5 -monophosphate (dAMP), 2 -deoxycytidine 5 -monophosphate(dCMP), 2 -deoxyguanosine 5 -monophosphate (dGMP), and2 -deoxythymidine 5 -monophosphate (d MP))24. Base selectivity was achieved by modi ying a mutant -haemolysin pore withaminocyclodextrin adapter covalently attached within the -barthereby constricting the channel while enhancing the chemispeci city o the sensor. Raw mononucleotides were read with o99% con dence under optimal operating conditions. By integratthis base identi cation plat orm with a highly processive exonuc

ase (through either chemical attachment or genetic usion) to cleand successively pass nucleotides to the nanopore, a single molesequencing by digestion approach may indeed be easible. OxNanopore echnologies is currently developing a DNA sequenctechnology based on this approach.

Although -haemolysin has dominated the biological nanopsequencing landscape so ar, other more e cient biological nanpores are emerging. A structural drawback with -haemolysinthat the cylindrical -barrel can accommodate up to ~10 nucltides at a time, all o which signi cantly modulate the pore curent25: this dilutes the ionic signature o the single nucleotide in 1.4 nm constriction, thus reducing the overall signal-to-noise rain sequencing applications. Te octameric protein channel MspA26 does not su er rom this problem as it has a channel shaped li

a tapered unnel that is just ~0.5 nm long, with a constrictiondiameter ~1.2 nm at the point where it meets the vestibule (Fig. Genetically engineered MspA can discriminate between

nucleotide sets (AAA, GGG, , CCC) with an impressive 3.5old enhancement in nucleotide separation e ciency over nativ-haemolysin27. Moreover, in experiments involving immobilizessDNA, as ew as three nucleotides within or near the constrtion contributed to the pore current27 compared with the ten or sonucleotides that modulate the current in native -haemolysin25. By using additional genetic techniques (notably site-speci c mutageesis) it might be possible to achieve single-nucleotide resolutionMspA-based sequencing also aces challenges. Te speed o unipeded ssDNA translocation through MspA is too ast to sequessDNA directly and in real time. Duplex interrupted nanopo

sequencing27

might be able to overcome this limitation by insertia short segment o dsDNA between each nucleotide in an anaDNA molecule. As the converted DNA is driven through an Mnanopore, each duplex temporarily halts the translocation proceallowing a single analyte nucleotide to be identi ed by measurthe ionic current. Te duplex would then dissociate owing to thhigh electric eld in the nanopore, allowing the DNA to advancethe next duplex and the next analyte nucleotide to be determinand so orth. Such a method could ultimately enable ast and losequential reads. However the ability to convert and read lagenomic ragments with high delity using this approach has nyet been demonstrated.

Strand sequencing is an alternative to duplex interruptsequencing that involves coupling an enzyme motor to a nanopthis makes it possible to controllably step ssDNA through Ms

106

105

s s D N A

( 1 5 ) 1

0 0 n t

( 5 6 )

3 k b

p

( 5 7 )

6 . 6 k

b p

( 1 0 9 )

3 k b

p

( 3 8 ,

4 5 ) 5

k b p

( 5 8 ) 4 0 0

b pIdeal

velocityrange

for

-haemolysin

( 1 0 7 )

3 k b

p

( 1 0 8 )

1 1 k b

p

( 1 1 0 )

5 3 9 b

p

( 1 1 1 )

1 5 0 b p

( 6 1 )

2 5 b p ( 1 1

2 ) 2 . 7 k

b p

Solid state

( 1 0 4 )

7 n t

( 1 6 ) 1

0 0 n t

( 1 0 3 ) 1 0 0 n

t

( 9 8 )

1 n t

( 1 0 6 )

3 9 n t

( 2 4 )

1 n t

( 2 1 )

7 9 n t

( 1 0 5 )

6 0 n t

104

103

102

101

100

101

1021995 2000 2005 2010

Year

D N A v e l o c

i t y

( n t m s

1 )

2015 2020

d s D N

A

Figure 1 | Trends in nanopore analysis o DNA. DNA translocation velocityin nucleotides per millisecond (on a logarithmic scale) versus year or bothssDNA and dsDNA and or both -haemolysin and solid-state nanopores.Each data point represents a reduction in translocation velocity or animprovement in sensitivity; or each data point the relevant re erence(s)are shown in parentheses and the size o the shortest molecule detected isshown in nucleotides (nt), base pairs (bp) or kilobase pairs (kbp). For solid-state nanopore sequencing applications, the translocation velocity shouldbe in the range 1100 nt ms 1 (pale blue region). Biological nanopores havealready reached velocities below this range, and solid-state nanopores arealso expected to approach these values around 2015. However, substantialimprovements in sensitivity o both biological and solid-state nanopores arealso needed, which is why researchers are exploring the new approachesdescribed in Table 1. The reduced velocities (and improved sensitivities)or -haemolysin have been achieved by a combination o site-speci cmutagenesis and one o the ollowing: the incorporation o DNA processingenzymes into the nanopore 21, chemical labelling o the nucleotides 106 or thecovalent attachment o an aminocyclodextrin adapter 24 . The improvements

in the per ormance o solid-state nanopores have been due to theoptimization o solution conditions (temperature, viscosity, pH) 109, chemicalunctionalization 110, sur ace-charge engineering 45 , varying the thicknessand composition o the membranes 38,61,112, and the use o smaller diameternanopores 58,61 (thereby enhancing polymerpore interactions). This gurecontains key nanopore developments in terms o translocation velocity andsensitivity, but is by no means exhaustive.

REVIEW ARTICLE NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.129

2011 Macmillan Publishers Limited. All rights reserved


3/10

NATURE NANOTECHNOLOGY | VOL 6 | OCTOBER 2011 | www.nature.com/naturenanotechnology 617

20 400

300

600

900

1,200

E v e n t c o u n t

Residual pore current (pA)

Residual pore current (pA)

dGMPdTMPdAMPdCMP

0 0.2 0.4Time (s)

R e s i

d u a

l p o r e

c u r r e n t

( p A )

10

20

30

40

50

60a

b

T T T TA A A C CG-barrel

1.4 nm

Vestibule

2.6 nmconstriction

1.2 nmconstriction

2 n m

Vestibule

GGG TTT CCC AAA

1 1.50

40

30

20

0.5

R e s i

d u a

l p o r e

c u r r e n t

( p A )

Time (s)40 50 60 70

N o r m a

l i z e

d c o u n t

GGG

CCCTTT

AAA50

Table 1 | Di erent sensing modalities or nanopore sensors and the challenges they ace.

Sensing modality Description o technique Potential challenges Re erences

Ionic current Hybridization-assisted nanopore sequencing High spat ial resolution required; complex algori thmsneeded or analysis.

97

Sequencing by exonuclease digestion Requires sequential passage o mononucleotides in theorder in which they are cleaved.

24, 98

Sequencing by synthesis Retaining processing enzymes (DNA polymerase) at the

pore; achieving long read lengths and maintaining enzymeactivity under a voltage load.

1821

Duplex interrupted DNA sequencing Converting large genomic ssDNA ragments to duplexinterrupted structure.

27

Optical readout Optical recognition o converted DNA Complex and error-prone DNA conversion steps; highdensity,


4/10


with nucleotide identi cation occurring at each step. Candidateenzymes suited or this application include 7 DNA polymerase,Klenow ragment o DNA polymerase 1 and phi29 DNA polymer-ase18,20,21. Te last o these is known to be remarkably stable andhighly e cient at catalysing sequential nucleotide additions in-haemolysin20. It is plausible, there ore, that long strands o DNAcould be sequenced by coupling phi29 DNA polymerase to MspA,although this has not yet been demonstrated experimentally.

Biological nanopores also have tremendous potential orapplications other than DNA sequencing. Te connector proteinrom the bacteriophage phi29 DNA packaging motor, or exam-ple, could prove use ul or applications in molecular diagnosticsand DNA ngerprinting. Te channel in this protein nanopore isapproximately 3.6 nm wide, and it opens into a vestibule with innerdiameter o ~6 nm. Moreover, it has an open channel conductancethat is approximately ve times higher than that o -haemolysin under similar conditions. Tis suggests the possibility o screeninglarger analytes such as dsDNA, DNA protein complexes and aminoacid polymers or protein sequencing. Te translocation o dsDNAthrough a genetically engineered connector channel embedded in alipid bilayer was recently demonstrated28. Unidirectional transporto dsDNA through this channel ( rom amino-terminal entrance tocarboxyl-terminal exit) was also observed29, suggesting a natural

valve mechanism in the channel that assists dsDNA packaging dur-ing bacteriophage phi29 virus maturation. Te capabilities o thisprotein nanopore will become more apparent in years to come.

Despite the heterogeneity and remarkable sensitivity o biologicalnanopores, they also have some inherent disadvantages: the mechani-cal instability o the lipid bilayer that supports the nanopore; the sen-sitivity o biological nanopores to experimental conditions (such aspH, temperature and salt concentration); and the di culty in inte-grating biological systems into large-scale arrays. Te stability o the

bilayer can be improved by supporting it on solid and nanoporsubstrates3033, by varying its compositions34,35 or by using tetheredbilayer architectures36. However, as we shall see in the next sectiosolid-state nanopores are considerably more robust and durable.

Solid-state nanoporesSolid-state nanopores are ast becoming an inexpensive and hig versatile alternative to biological nanopores. As well as robusand durability, the solid-state approach o ers the ability to tune tsize and shape o the nanopore with subnanometre precision37,38,the ability to abricate high-density arrays o nanopores39, superiormechanical, chemical and thermal characteristics compared wlipid-based systems, and the possibility o integrating with etronic40 or optical readout techniques41.

Te rst reports o DNA sensing using solid-state nanoporeemerged in early 2001 when Golovchenko and co-workers uscustom-built ion-beam sculpting tool with eedback control to mnanopores with well-de ned sizes in thin SiN membranes42. Now,most groups pre er to use a ocused electron beam with a probe tis a ew nanometres in diameter to sputter nanopores in thin insuling membranes, a technique that has evolved since the 1980s43. Teabrication o solid-state nanopores in thin insulating membranes hbeen reviewed elsewhere44, as have applications to single-molecul

biophysics10. SiN has traditionally been the nanopore membranmaterial o choice owing to its high chemical resistance and mechanical stress. Tese membranes have traditionally been aricated by an optimized low-pressure chemical vapour depositprocess, typically at elevated temperatures (~800 C), but this prodoes not o er the subnanometre control over membrane thicknethat is needed to probe DNA with single-nucleotide resolution.

In working towards developing ultrathin membranes, one o(R.B.) and co-workers have shown that atomic-layer deposi

_

a b c

+

4

40

t = 0 mins t = 4 mins t = 6 mins

E v e n t

b l o c k a g e

( % )

20

0.01 0.1 1 10Event duration (ms)

100

3

2

1

0

E v e n t

b l o c k a g e

( n A )

4002000Event duration ( s)

1 . 3 n

A

Constant ecd

200 s

10 nm

5 nm 5 nm 5 nm

5 nm

Figure 3 | Solid-state nanopore architectures or DNA analysis. a , Top: TEM images showing the ormation (by a ocused electron beam) andcontrolled contraction o nanopores in Al 2O3 membranes. It is possible to control the diameter o the nanopore with subnanometre precision. Bottom:scatter plot o event blockage (the percentage o open pore ionic current that is blocked as a molecule passes through the pore) versus event duration(on a logarithmic scale) or the translocation o segments o dsDNA containing 5,000 base pairs through a 5-nm-diameter Al 2O 3 pore showing a singleblockage level (~20% o the open pore current) corresponding to linear, un olded dsDNA transport. b , Top: schematic (le t) and TEM image (right)o a nanopore in a suspended graphene lm containing just one to two layers o carbon atoms. Bottom: scatter plot o event blockage versus eventduration showing that olded DNA (le t o inset, deep blockade level) and un olded DNA (right o inset, shallow blockade level) can be distinguished.The solid line represents a constant event charge de cit (ecd; which is the total area under each current transient resulting rom a DNA translocationevent). As shown, ast high-amplitude events and slow low-amplitude events can share a constant ecd or constant area, signi ying olded and un oldedtranslocation events respectively. c, Top: TEM image o a terraced nanopore ormed in a graphene lm containing ~10 monolayers o carbon atoms.Scale bar, 1 nm. Bottom le t: nanopore in a monolayer o graphene with primarily armchair edges surrounded by multilayered regions. Scale bar, 1 nm.Bottom right: TEM image o a nanopore in multilayer graphene; ripples at the pore edge again show the terraced structure. Figures reproduced withpermission rom: a , re . 38, 2009 Wiley and re . 45, 2010 Wiley (inset); b, re . 51, 2010 NPG; c, re . 55, 2011 ACS (top and bottom le t panels)and re . 52, 2010 ACS (bottom right panel).




5/10


(ALD) can be used to make Al2O3 membranes, with ngstrm-levelcontrol over membrane thickness38, and that two interesting phenom-ena occur when a ocused electron beam is used to sputter nanoporesin these metal oxide membranes. First, the Al2O3, which is an insula-tor, is partially converted into metallic Al, in a dose-dependent way this process could potentially be used to write nanoscale electrodesin the pore. Second, di erent nanocrystalline domains are ormedin a dose-dependent manner, which could potentially allow pattern-ing o the sur ace charge at the inter ace between the nanopore andthe electrolyte45. Controlling the stoichiometry o the material in thenanopore and/or the sur ace charge density could allow researchersto reduce 1/ noise46 and DNA translocation velocities. Indeed, weobserved that DNA translocation was slower in Al2O3 nanoporesthan in SiN nanopores with similar diameters, which was attributed

to the strong electrostatic interactions between the positively chargedAl2O3 sur ace and the negatively charged dsDNA45. Enhancing theseinteractions, either electrostatically or chemically, could reduce DNA velocities even more. Te ALD approach should also allow nanoporesto be ormed in a variety o other high-dielectric-constant mate-rials, including iO2 and H O2, each o which has unique materialproperties. Moreover, ALD is a relatively low-temperature (typically


6/10


However, Dekker and co-workers53 ound that the pore conduct-ance was proportional to the square o the diameter, which sug-gests that the thickness o the membrane was not negligible (thatis,Rpore > Raccess). Te origin o this behaviour might be the polymercoating (16-mercaptohexadecanoic acid) that was used to coat thegraphene to reduce DNA adsorption. Tey also ound that the con-

ductance or a given diameter o nanopore remained largely constantas the number o graphene layers in the membrane was increasedrom one to eight53. Tis result is plausible as nanopore ormationin multilayer graphene is known to induce a terrace e ect, with thenumber o layers increasing as we move away rom the pore (Fig. 3c).Tis e ect was con rmed recently using EM image analysis55, and isalso visible in earlier studies49. A terraced nanopore architecture couldprove very use ul or two reasons: (1) it potentially relaxes the con-straint o growing and trans erring a large-area monolayer to abricatea graphene monolayer nanopore; and (2) a multilayered support may increase the stability and longevity o a graphene nanopore sensor.

Te translocation o dsDNA through graphene nanoporesinduced subtle uctuations in the ionic current, indicating the trans-location o both olded and un olded DNA structures5153, analogous

to the DNA-induced current blockades seen in SiN nanopores56,57

.ranslocation velocities were between about 10 and 100 nucleotidesper microsecond, which is too ast or the electronic measuremento individual nucleotides. However, computer simulations o dsDNApassing through a 2.4-nm-diameter nanopore in a ~0.6-nm-thickgraphene membrane revealed a resolution o ~0.35 nm, which issimilar to the size o an individual DNA nucleotide51. Tis resultsuggests that i the translocation speed could be reduced to roughly one nucleotide per millisecond, single-nucleotide detection should bepossible, which could potentially lead to DNA sequencing with elec-tronic readout.

o enable such advancements, however, we need a better under-standing o the quantitative aspects o DNA translocation. Why, orexample, do nanopores in multilayer graphene (315 monolayers)52 give deeper DNA-induced current blockades than nanopores in

single-layer graphene under normalized conditions51. One possibleexplanation is the terrace e ect described above, but the structure ographene nanopores needs to be studied in greater detail.

Tere are also various undamental questions about sequencing with graphene nanopores that remain unanswered. For exaple, is single-nucleotide resolution possible in the presence

thermodynamic uctuations and electrical noise? And will tchemical and structural similarity o the purines (A and G) athe pyrimidines (C and ) inherently limit the identi cation oindividual nucleotides using ionic current? Sur ace unctioalization might enhance the nucleotide speci city o graphenanopores, but it could also compromise resolution by increasthe thickness o the membrane. Without doubt, this exciting ypreliminary work will certainly be the precursor or many utustudies o graphene nanopores.

Other applications o nanoporesTe more immediate application or solid-state nanopores is likelybe in medical diagnostics. A nanopore-based diagnostic tool coo er various advantages: it could detect target molecules at very

concentrations rom very small sample volumes58

; it could simultane-ously screen panels o biomarkers or genes (which is importandisease diagnosis, monitoring progression and prognosis); it coprovide rapid analysis at relatively low cost; and it could elimicumbersome ampli cation and conversion steps such as PCR, bisphite conversion and Sanger sequencing.

MicroRNA (miRNA) expression pro ling is one applicatiwhere solid-state nanopore technology could excel. Te detetion and accurate quanti cation o these cancer biomarkers whave important clinical implications, acilitating disease diagnostaging, progression, prognosis and treatment response59,60. It wasrecently demonstrated that a nanopore-based approach or thdetection o speci c microRNA sequences enriched rom cellutissue can achieve sensitivities that are comparable to conventimicroarray technologies (Fig. 4a)61 .

ssDNA

Si

SiO22

12

9

6

3

01.0 1.5 2.0 2.5

Time (s)

C o n

d u c t a n c e

( n

S )

100

200

C o u n t

Event blockage (nA)

Streptavidin Anti-biotin antibodies

0

50

100

00 0.8 1.6

10 nm

1 MM DNA

PC DNA

b ca

2,000

E v e n t

d u r a t i o n

( m s )

1,500

1,000

500

00 100 200 300

Event blockage (pA)400 500

Si

DNA hairpins~ 12 nm~ 1618 nm

SiO2

10 nm

0 0.8 1.6

10 nm

Figure 5 | Hybrid biologicalsolid-state nanopores. a , A SiO2 nanopore unctionalized with hairpin DNA (top) can distinguish between ssDNA that isper ectly complementary (PC DNA) to the hairpin sequence (bottom, red data points) and ssDNA that di ers by a single base (bottom, blue data points;1MM, one base mismatch), thereby allowing or the detection o SNPs. b, A SiN nanopore coated with a fuid lipid bilayer containing mobile ligandsattached to the sur ace (top) can detect and di erentiate various protein analytes because they lead to di erent current blockage distributions as shownin the histograms (bottom). c, -haemolysin can be inserted into a SiN nanopore by attaching a dsDNA tail that can be pulled through the nanoporeby electrophoresis (top le t). The ormation o the hybrid nanopore occurs in three stages (top right), each o which is associated with a characteristicconductance (bottom). When the -haemolysin has been inserted into the SiN nanopore (stage III), the conductance is consistent with values measuredor -haemolysin in a lipid bilayer. Figures reproduced rom: a , re . 70, 2007 NPG; b, re . 74, 2011 NPG; c, re . 75, 2010 NPG.




7/10


Another potential application or solid-state nanopores isepigenetic analysis more speci cally the detection o aberrantDNA methylation, which can serve as a robust biomarker in cancer62 ( or example, GS P1 promoter hypermethylation is observed inover 90% o prostate cancer cases)63. Aberrant methylation can alsobe an indicator o disease severity and metastatic potential in many tumour types62,64. Preliminary progress towards nanopore-basedmethylation analysis has been reported, including the detection o methylated65 and hydroxymethylated DNA66, but this application isstill in its in ancy.Genetic analysis involving the detection o single nucleotidepolymorphisms (SNPs) is another important diagnostic applicationor which nanopores may be well suited. SNPs and point mutationshave been linked to a variety o mendelian diseases as well as morecomplex disease phenotypes67. In proo -o -principle experiments,SNPs have been detected using ~2-nm-diameter SiN nanopores68.Using the nanopore as a local orce actuator, the binding energies o a DNA binding protein and its cognate sequence relative to a SNPsequence could be discriminated (Fig. 4b). Tis approach could beextended to screen mutations in the cognate sequences o variousother DNA binding proteins, including transcription actors, nucle-ases and histones.

Te genomic pro ling o viruses and human pathogens using

solid-state nanopores is also attractive. An innovative approachinvolving the introduction o highly invasive peptide nucleic acid(PNA) probes has been used to label target genomes with high a n-ity and sequence speci city, creating local bulges (P loops) in themolecule69. ranslocation o this labelled molecule resulted in sec-ondary DNAPNA blockade levels (Fig. 4c), e ectively barcodinga target genome. Although urther studies are needed to determinethe ultimate spatial resolution o this technique, it could potentially enable the rapid, accurate and ampli cation- ree identi cation o small (510 kilobase) viral genomes including hepatitis C, dengueever and West Nile virus.

Hybrid biologicalsolid-state nanoporesA major drawback with solid-state nanopore technology at pre-

sent is that it cannot chemically di erentiate between analytes o approximately the same size. Tis lack o chemical speci city can beovercome by attaching speci c recognition sequences and receptorsto the nanopore to create a hybrid structure70,71that has the potentialto uniquely identi y nucleotides in sequencing applications, or todi erentiate and quanti y target proteins in diagnostic applications.Chemical unctionalization and its e ect on the electrical propertieso polymer nanopores was recently reviewed72.

Sur ace unctionalization can also be used to introduce DNAsequence speci city. In studies involving SiO2 nanopores unction-alized with hairpin DNAs, higher uxes and shorter translocationtimes were observed or the passage o per ectly complementary ssDNA compared with ssDNA in which there was a single basemismatch, demonstrating the potential o this approach to detect

SNPs (Fig. 5a)70

. Functionalized biomimetic nanopores in SiN haveurthermore enabled the study o nucleocytoplasmic transport phe-nomena at the single-molecule level73.

Altering the sur ace chemistry o a nanopore can also acilitate thesensitive detection and discrimination o proteins. Drawing inspira-tion rom the lipid-coated ol actory sensilla o insect antennae, SiNnanopores coated with a uid lipid bilayer were recently used to iden-ti y proteins (Fig. 5b)74. Te incorporation o mobile ligands into thebilayer introduced chemical speci city into the nanopore, slowedthe translocation o target proteins, prevented pores rom cloggingand eliminated non-speci c binding, thereby resolving many issuesinherent to solid-state nanopores. A lipid-bilayer-coated nanoporearchitecture o this nature (in either SiN (re . 74) or Al2O3 (re . 33))could also, in principle, be integrated with biological nanopores toorm robust nanopore sequencing elements.

Te concept o hybrid biologicalsolid-state nanopores tooanother step orward recently when a genetically engineer-haemolysin protein channel was inserted into nanopore in a Smembrane (Fig. 5c)75. By chemically linking a long dsDNA tail -haemolysin, it could be electrophoretically guided into the nanopore to orm a coaxially aligned structure. Measuremeno nanopore conductance and ssDNA translocation times wein good agreement with -haemolysin embedded in lipid bilay-ers76, but blockade amplitudes were signi cantly less, which wattributed to de ormation o the biological nanopore and leakacurrents, and an increase in electrical noise was also observTese parameters will probably need to be optimized to matc

a

b

A

Figure 6 | Possible novel nanopore architectures or sequencing.a , Schematic cross-section showing ssDNA passing through a solid-statenanopore with an embedded tunnelling detector 92 . The detector consistso two electrodes spaced ~1 nm apart on opposite sides o the nanopore.

Changes in the tunnelling current as ssDNA passes through the nanopore(and between the electrodes) could be used to identi y the sequenceo bases in the DNA. Inset: top view showing a nucleotide positioned inthe nanogap between the electrodes o the tunnelling detector. Figurereproduced rom re . 92, 2010 NPG. b, Schematic showing ssDNApassing through a solid-state nanopore with an embedded graphene gateand a graphene nanoribbon on the membrane containing the nanopore;both the gate and the nanoribbon contain circular openings or the DNAto pass through. The graphene gate could be used to induce either p-typeor n-type behaviour in the nanoribbon and to electrostatically slowdown ssDNA. Changes in the transverse conductance o the nanoribbonas ssDNA passes through the nanopore could be used to identi y thesequence o bases in the DNA 84 . Functionalization o the edges o thecircular opening in the nanoribbon could urther enhance nucleotide-

speci c interactions.

REVIEW ARTICLENATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.129



8/10


the single nucleotide sensitivity o aminocyclodextrin-modi ed-haemolysin24. Nevertheless, this hybrid architecture opens up theexciting possibility o combining the single-nucleotide recognitioncapabilities o -haemolysin or MspA with wa er-scale arrays o individually addressed solid-state nanopores or high-throughputsequencing applications.

OutlookTe advances described here suggest that nanopores are likely tohave an increasing role in medical diagnostics and DNA sequenc-ing in years to come, but they will ace competition rom a numbero other techniques. Tese include single-molecule evanescent elddetection o sequencing-by-synthesis in arrays o nanochambers(Paci c Biosciences)77, sequencing by ligation on sel -assembled DNAnanoarrays (Complete Genomics)78, and the detection o H+ ionsreleased during sequencing-by-synthesis on silicon eld-e ect tran-sistors rom multiple polymerase-template reactions (Ion orrent)79.However, the possibility o using nanopore-based sensors to per ormlong base reads on unlabelled ssDNA molecules in a rapid and cost-e ective manner could revolutionize genomics and personalizedmedicine, which is why public bodies such as the National Instituteso Health80 in the US and private-sector companies such as IBM,Ox ord Nanopore echnologies and NABsys, are continuing to invest

heavily in nanopore-based DNA sequencing technology.Current trends suggest that many challenges in sequencingwith biological nanopores (the high translocation velocity and thelack o nucleotide speci city) have been resolved. Similarly, giventhe progress with solid-state nanopores, i the translocation veloc-ity could be reduced to a single nucleotide (which is ~3 long)per millisecond, and i nucleotides could be identi ed uniquely with an electronic signature (an area o intense research), it wouldbe possible to sequence a molecule containing one million basesin less than 20 minutes. Furthermore, i this technology could bescaled to an array o 100,000 individually addressed nanoporesoperating in parallel, it would be possible to sequence an entirehuman genome (some three billion base pairs) with 50- old cover-age in less than one hour.

Achieving this goal will require novel device architectures thatcomplement or replace ionic current readout with more sophisti-cated readout methods. Tere have been preliminary reports on theuse o embedded planar gate electrodes in nanopores40 and nano-channels81,82to electrically modulate the ionic pore current, and theintegration o single-walled carbon nanotubes or the translocationo ssDNA83. Teorists, meanwhile, have proposed embedding gra-phene nanoribbons84,85 and nanogaps86,87 in nanopores to enableelectronic readout o individual nucleotides, but issues such as theelectrochemistry at the graphene/ uid inter ace, the reproducibility o the electronic properties o the graphene edges and the chemicalunctionalization o these edges need to be addressed. An alterna-tive architecture, termed the DNA nanopore transistor, has beenproposed by IBM researchers. Using molecular dynamics, the IBM

team recently demonstrated the controlled base-by-base ratchetingo ssDNA through a nanopore in a multilayer metal oxide mem-brane (driven by alternating electric elds applied across the metallayers)88, but this result has yet to be con rmed in experiments.

Recent experiments with scanning tunnelling microscopes sug-gest that it might be possible to identi y nucleotides with electrontunnelling89 (because the energy gaps between the highest occu-pied and lowest unoccupied molecular orbitals o A, C, G and areunique90), and partially sequence DNA oligomers91. Nano abricatedmetallic gap junctions have also been used to identi y single nucleo-tides by measuring electron tunnelling currents92. Neither o theseexperiments involved nanopores, but i a tunnelling detector o thisnature could be embedded in a solid-state nanopore, and the DNAtranslocation velocity could reduced, it may become possible todirectly sequence DNA with a solid-state nanopore (Fig. 6).

E orts to abricate nanopore sensors that contain nanogap-basetunnelling detectors are currently underway 93,94, but thermal uc-tuations and electrical noise present major challenges. One solutcould be to adopt a statistical approach that involves repeatedly spling each nucleotide or molecule to determine the sequence o baAnother challenge is the act that tunnelling currents vary expontially with both the width and the height o the barriers that electrhave to tunnel through, which in turn depends on the e ective tunndistance and on molecule orientation. A our-point-probe measument could there ore reveal signi cantly more in ormation than thtwo-probe measurements attempted so ar, but reliably abricatisuch a our-probe structure with subnanometre precision will a ormidable challenge. It should also be noted that it is not necsary to uniquely identi y all our bases or certain applications. Soresearchers have used a binary conversion o nucleotide sequen(A or = 0, and G or C = 1), to discover biomarkers and identgenomic alterations in short ragments o DNA and RNA95,96.

Over the past ew years both biological and solid-state nanopohave been moving closer to the goal o direct label- ree sequenco DNA molecules in real time. Tere is no doubt that nanoporbased sensors will continue to develop as strong candidates to other third-generation sequencing technologies in the race towaa ordable and personalized DNA sequencing.

Re erences1. Tomas, P. D. & Kejariwal, A. Coding single-nucleotide polymorphisms

associated with complex vs. mendelian disease: Evolutionary evidenceor di erences in molecular e ects.Proc. Natl Acad. Sci. USA

101, 1539815403 (2004).2. International HapMap Consortium. A haplotype map o the human genome

Nature 437, 12991320 (2005).3. Mardis, E. R. Next-generation DNA sequencing methods.Annu. Rev. Genom.

Hum. Genet. 9, 387402 (2008).4. Metzker, M. L. Sequencing technologies the next generation.Nature Rev.

Genet. 11, 3146 (2010).5. http://www.genome.gov/12513210.6. Coulter, W. H. Means or counting particles suspended in a uid. US patent

2,656,508 (1953).7. Church, G., Deamer, D. W., Branton, D., Baldarelli, R. & Kasianowicz, J.

Characterization o individual polymer molecules based on monomer-inter ace interactions. US patent 5,795,782 (1995).

8. Deamer, D. W. & Branton, D. Characterization o nucleic acids by nanoporanalysis. Acc. Chem. Res. 35, 817825 (2002).

9. Rhee, M. & Burns, M. A. Nanopore sequencing technology: research trenand applications.Trends Biotechnol. 24, 580586 (2006).

10. Dekker, C. Solid-state nanopores.Nature Nanotech. 2, 209215 (2007).11. Healy, K. Nanopore-based single-molecule DNA analysis.Nanomedicine

2, 459481 (2007).12. Branton, D.et al. Te potential and challenges o nanopore sequencing.

Nature Biotechnol. 26, 11461153 (2008).Tis review article assesses the easibility o various nanopore sequencing techniques that are currently under development (both optical and electrical ).

13. Deamer, D. W. Nanopore analysis o nucleic acids bound to exonucleases polymerases. Annu. Rev. Biophys. 39, 7990 (2010).

Tis review article provides a historical perspective on the eld o nanopore DNA sequencing and elaborates on nanopore -based enzyme -mediated sequencing approaches .

14. Iqbal, S. & Bashir, R.Nanopores: Sensing and Fundamental Biological Interactions (Springer, 2011).

15. Kasianowicz, J. J., Brandin, E., Branton, D. & Deamer, D. W. Characterizao individual polynucleotide molecules using a membrane channel.Proc. Natl Acad. Sci. USA 93, 1377013773 (1996).Tis is the rst experimental report o electronic detection o a DNA molecule passing through a nanopore ( -haemolysin ) and marks the start o the nanopore sequencing eld .

16. Akeson, M., Branton, D., Kasianowicz, J. J., Brandin, E. & Deamer, D. WMicrosecond time-scale discrimination among polycytidylic acid,polyadenylic acid, and polyuridylic acid as homopolymers or as segmentswithin single RNA molecules.Biophys. J. 77, 32273233 (1999).

17. Meller, A. & Branton, D. Single molecule measurements o DNA transporthrough a nanopore.Electrophoresis 23, 25832591 (2002).




9/10


10/10


77. Eid, J.et al. Real-time DNA sequencing rom single polymerase molecules.Science 323, 133138 (2009).

78. Drmanac, R.et al. Human genome sequencing using unchained base reads onsel -assembling DNA nanoarrays.Science 327, 7881 (2010).

79. Rothberg, J. M.et al. An integrated semiconductor device enabling non-optical genome sequencing.Nature 475, 348352 (2011).

80. http://www.genome.gov/27527584.81. Karnik, R., Duan, C., Castelino, K., Daiguji, H. & Majumdar, A. Recti cation

o ionic current in a nano uidic diode.Nano Lett. 7, 547551 (2007).82. Jin, X. & Aluru, N. R. Gated transport in nano uidic devices. Microfuid.

Nanofuid. 11, 297306 (2011).83. Liu, H.et al. ranslocation o single-stranded DNA through single-walledcarbon nanotubes.Science 327, 6467 (2010).

84. Nelson, ., Zhang, B. & Prezhdo, O. V. Detection o nucleic acids withgraphene nanopores:ab initio characterization o a novel sequencing device.Nano Lett. 10, 32373242 (2010).

85. Min, S. K., Kim, W. Y., Cho, Y. & Kim, K. S. Fast DNA sequencing with agraphene-based nanochannel device.Nature Nanotech. 6, 162165 (2011).

86. Postma, H. W. C. Rapid sequencing o individual DNA molecules in graphenenanogaps.Nano Lett. 10, 420425 (2010).

87. Prasongkit, J., Grigoriev, A., Pathak, B., Ahuja, R. & Scheicher, R. H.ransverse conductance o DNA nucleotides in a graphene nanogap rom rstprinciples.Nano Lett. 11, 19411945 (2011).

88. Luan, B.et al. Base-by-base ratcheting o single stranded DNA through asolid-state nanopore.Phys. Rev. Lett. 104, 238103 (2010).IBMs DNA transistor architecture and proposed approach to nanopore -based single -molecule DNA sequencing are presented here .

89. Huang, S.et al. Identi ying single bases in a DNA oligomer with electrontunnelling.Nature Nanotech. 5, 868873 (2010).

90. Zwolak, M. & Di Ventra, M. Physical approaches to DNA sequencing anddetection.Rev. Mod. Phys. 80, 141165 (2008).

91. anaka, H. & Kawai, . Partial sequencing o a single DNA molecule with ascanning tunnelling microscope.Nature Nanotech. 4, 518522 (2009).

92. sutsui, M., aniguchi, M., Yokota, K. & Kawai, . Identi ying singlenucleotides by tunnelling current.Nature Nanotech. 5, 286290 (2010).Tis is the rst report o a nano abricated gap junction being used to discriminate individual nucleotides through electron tunnelling measurements .

93. aniguchi, M., sutsui, M., Yokota, K. & Kawai, . Fabrication o the gatingnanopore device. Appl. Phys. Lett. 95, 123701123703 (2009).

94. Ivanov, A. P.et al. DNA tunneling detector embedded in a nanopore.NanoLett. 11, 279285 (2010).

95. Asmann, Y. W., Kosari, F., Wang, K., Cheville, J. C. & Vasmatzis, G.

Identi cation o di erentially expressed genes in normal and malignantprostate by electronic pro ling o expressed sequence tags.Cancer Res. 62, 33083314 (2002).

96. Feldman, A. L.et al. Discovery o recurrent t(6;7)(p25.3;q32.3) translocationsin ALK-negative anaplastic large cell lymphomas by massively parallelgenomic sequencing.Blood 117, 915919 (2010).

97. Ling, X. S., Bready, B. & Pertsinidis, A. Hybridization-assisted nanoporesequencing o nucleic acids. US patent 2007 0190542 (2007).Tis approach , called HANS , is being commercially developed by NABsys:6-mer oligonucleotide probes are hybridized to ssDNA and current versustime traces reveal the position o the probe on the ssDNA template .

98. Astier, Y., Braha, O. & Bayley, H. oward single moleculeDNA sequencing: direct identi cation o ribonucleoside anddeoxyribonucleoside 5 -monophosphates by using an engineered proteinnanopore equipped with a molecular adapter. J. Am. Chem. Soc. 128, 17051710 (2006).

99. Lagerqvist, J., Zwolak, M. & Di Ventra, M. Fast DNA sequencing via tranelectronic transport.Nano Lett. 6, 779782 (2006).

100. Heng, J. B.et al. Beyond the gene chip.Bell Labs Tech. J. 10, 522 (2005).101. Gracheva, M. E.et al. Simulation o the electric response o DNA

translocation through a semiconductor nanopore-capacitor.Nanotechnology

17, 622633 (2006).102. Sigalov, G., Comer, J., imp, G. & Aksimentiev, A. Detection o DNAsequences using an alternating electric eld in a nanopore capacitor.Nano Lett. 8, 5663 (2007).

103. Meller, A., Nivon, L., Brandin, E., Golovchenko, J. & Branton, D. Rapidnanopore discrimination between single polynucleotide molecules.Proc. Natl Acad. Sci. USA 97, 10791084 (2000).

104. Howorka, S., Cheley, S. & Bayley, H. Sequence-speci c detection o individual DNA strands using engineered nanopores.Nature Biotechnol. 19, 636639 (2001).

105. Bates, M., Burns, M. & Meller, A. Dynamics o DNA molecules in amembrane channel probed by active control techniques.Biophys. J. 84, 23662372 (2003).

106. Borsenberger, V., Mitchell, N. & Howorka, S. Chemically labeled nucleotand oligonucleotides encode DNA or sensing with nanopores. J. Am. Chem.Soc. 131, 75307531 (2009).

107. Chen, P.et al. Probing single DNA molecule transport using abricatednanopores.Nano Lett. 4, 22932298 (2004).

108. Storm, A. J., Chen, J. H., Zandbergen, H. W. & Dekker, C. ranslocationo double-strand DNA through a silicon oxide nanopore.Phys. Rev. E 71, 051903 (2005).

109. Fologea, D., Uplinger, J., Tomas, B., McNabb, D. S. & Li, J. Slowing DNAtranslocation in a solid-state nanopore.Nano Lett. 5, 17341737 (2005).

110. Kim, Y. R.et al. Nanopore sensor or ast label- ree detection o short double-stranded DNAs.Biosensors Bioelectron. 22, 29262931 (2007).

111. Wanunu, M., Sutin, J., McNally, B., Chow, A. & Meller, A. DNAtranslocation governed by interactions with solid-state nanopores.Biophys. J. 95, 47164725 (2008).

112. Chen, Z.et al. DNA translocation through an array o kinked nanopores.Nature Mater. 9, 667675 (2010).

AcknowledgementsB.M.V. is a trainee supported by the Midwestern Cancer Nanotechnology rainingCenter (NIH-NCI R25 CA154015). Support rom the National Institutes o Health (RCA155863) and the National Science Foundation (EEC-0425626) is also acknowledTe authors thank J. Hanlon-Sinn (Beckman Institute o Advanced echnology,University o Illinois at Urbana-Champaign) or the images in Fig. 6, and M. Drndic(University o Pennsylvania) or valuable discussions.

Additional in ormationTe authors declare no competing nancial interests.


bioe 150 article 3

Documents