A molecular-imprint nanosensor for ultrasensitivedetection of proteinsDong Cai1*, Lu Ren1,4, Huaizhou Zhao2, Chenjia Xu1, Lu Zhang1, Ying Yu2,4, Hengzhi Wang2,

Yucheng Lan2, Mary F. Roberts3, Jeffrey H. Chuang1, Michael J. Naughton2, Zhifeng Ren2

and Thomas C. Chiles1

Molecular imprinting is a technique for preparing polymerscaffolds that function as synthetic receptors1–3. Imprintedpolymers that can selectively bind organic compounds haveproven useful in sensor development2–7. Although creatingsynthetic molecular-imprinting polymers that recognize pro-teins remains challenging8–11, nanodevices and nanomaterialsshow promise in this area12–14. Here, we show that arrays ofcarbon-nanotube tips with an imprinted non-conductingpolymer coating can recognize proteins with subpicogram perlitre sensitivity using electrochemical impedance spectroscopy.We have developed molecular-imprinting sensors specific forhuman ferritin and human papillomavirus derived E7 protein.The molecular-imprinting-based nanosensor can also discrimi-nate between Ca21-induced conformational changes in calmo-dulin. This ultrasensitive, label-free electrochemical detectionof proteins offers an alternative to biosensors based onbiomolecule recognition.

Molecular imprinting (MI) technology offers considerablepotential as a cost-effective alternative to the use of biomolecule-based recognition in a variety of sensor applications15–17.Molecular imprinting polymers (MIPs) afford the creation ofspecific recognition sites in synthetic polymers by a process thatinvolves co-polymerization of functional monomers and crosslin-kers around template molecules. The molecules are removed fromthe polymer, rendering complementary binding sites capable ofsubsequent template molecule recognition1–3. Although depositionof MIPs onto the surface of nanostructures may improve sensitivityfor the recognition of a range of organic compounds8,18–20, elec-tronic nanosensors capable of recognizing proteins continue to bea challenge to implement, in part because (i) the MIP film mayattenuate signals generated in response to template binding (dueto the large thickness); (ii) the detection mechanisms do notreadily allow for effective signal conversion of template moleculebinding; and (iii) the sensor platforms do not support highly sensi-tive detection4,9–11.

We have sought to overcome these limitations by imprinting anon-conducting polymer nanocoating on the tips of nanotubearrays. The template protein was initially incorporated into thenanocoating and, upon extraction of protein from the accessible sur-faces on the nanocoating, sensor electrical impedance was found tobe greatly reduced due to electrical leakage through the surfaceimprints in the nanocoating. Subsequent recognition of the templateprotein was detected as an increase in impedance due to the relativelylower conductivity of the protein. A critical component for the assem-bly of the sensor architecture comprised electropolymerizing a non-conductive polyphenol (PPn) nanocoating onto the tips of nanotubes

(Fig. 1a). The deposition was self-limiting (Supplementary Fig. S1)and yielded a highly conformal nanocoating that is beneficial inlow-noise recordings21.

Electrochemical impedance spectroscopy (EIS) revealed that thehighest impedance in the PPn coating occurred at the nanotube tips(data not shown), probably because of the fact that there is fasterelectron transfer on the nanotube tips than on the nanotube side-walls22. The nanotube tips have open cross-sections with centredcavities (Fig. 1b,c). Transmission electron microscopy (TEM) con-firmed that a 13-nm PPn thin film was uniformly deposited onthe nanotube tips and co-deposition of human ferritin (hFtn) wasvisualized by TEM due to contrast enhancement by the iron crystal-line cores in the hFtn (Fig. 1c). In general, the thickness of the PPnnanocoating was found to be comparable to the protein diameter, sothe impedance change in response to template protein bindingwould be significantly enhanced.

Estimation of the number of hFtn imprints was conducted byelectrochemically refilling the imprint voids with PPn, convertingthe total refilling charge to volume of PPn (ref. 23), and calculatingthe equivalent number of hFtn that represented the number ofimprints on the sensor (see Supplementary Information for calcu-lations). The average number of hFtn imprints per nanotube tipwas 12 (independent experiment with three chips). We also evalu-ated imprint development using streptavidin as a template molecule.Following rebinding of streptavidin, biotinylated gold particles wereused to detect the streptavidin bound on the nanotube tips(Supplementary Fig. S2), and the average number of streptavidinimprints per nanotube tip was found to be 40.

The detection of hFtn binding to its imprint site was evaluatedusing EIS and differential pulse voltammetry (DPV). Nyquistplots (Fig. 2a) show the impedance spectroscopy of the nanosensorat different stages of development and at various levels of proteinrebinding. Compared to nanotube tips devoid of PPn, it wasfound that the PPn coating with hFtn increased the sensor impe-dance modulus from 13+2 kV to 241+47 kV at f¼ 10 Hz. Eachmeasurement was preceded by measurement for a control protein(bovine serum albumin, BSA), which served as a reference for theresponse to hFtn. Adding hFtn in concentrations ranging from1 × 10212 to 1 × 1027 g l21 revealed a concentration-dependentincrease in impedance, which was significantly larger than thatinduced by BSA, even at 1 × 1023 g l21. The impedance modulusat 10 Hz indicated that the impedance change in response to hFtnoccurred at a concentration of 1 × 10211 g l21, whereas BSA didnot exert an impedance change below 1 × 1024 g l21. The impe-dance approached its maximum value at 1 × 1027 g l21 of hFtn.Thus, the dynamic range of hFtn detection spans four decades.

1Department of Biology, Boston College, Chestnut Hill, Massachusetts 02467, USA, 2Department of Physics, Boston College, Chestnut Hill, Massachusetts02467, USA, 3Department of Chemistry, Boston College, Chestnut Hill, Massachusetts 02467, USA, 4Institute of Nanoscience and Nanotechnology, CentralChina Normal University, Wuhan, China 430079. *e-mail: caid@bc.edu

LETTERSPUBLISHED ONLINE: 27 JUNE 2010 | DOI: 10.1038/NNANO.2010.114

NATURE NANOTECHNOLOGY | VOL 5 | AUGUST 2010 | www.nature.com/naturenanotechnology 597

© 2010 Macmillan Publishers Limited. All rights reserved.

The dissociation constant Kd of hFtn binding was 53.6 pg l21

(Supplementary Fig. S3). It is known that at low concentrations(for example, attomolar levels), the probability for a templatemolecule to bind to the imprint is no longer governed by simplediffusion rules, so the reproducibility of detection may becomelimited, potentially yielding false-negative responses. That said, asecond independent series of measurements were conducted usingDPV, with the peak current decreasing in response to increasinghFtn concentrations (Fig. 2b); the reduction in peak currentresponses was similar to that observed by EIS (Fig. 2c).

Horse apoferritin (ahsFtn) and horse ferritin (hsFtn) did notproduce a significant response with the hFtn-imprinted nanosensorwhen evaluated at concentrations as high as 1 × 1024 g l21

(Fig. 2d). The specificity of the hFtn-imprinted nanosensor wasalso demonstrated in a series of measurements with hFtn combinedwith ahsFtn or hsFtn (Fig. 2d). In each case, binding of hFtn wasachieved, as indicated by a decrease in the percentage of peakcurrent, with each binary mixture showing decreased DPVresponses greater than the ahsFtn or hsFtn alone. Similar resultswere obtained with a complex mixture of proteins derived frombovine muscle extracts (Fig. 2d).

To investigate the mechanism of detection following templateprotein binding, Nyquist plots were fitted with a model containingconstant phase elements (CPEs) (Fig. 3a). In Fig. 3a, Rp is the PPncoating resistance and Ru the solution resistance. When a0 in CPE0is �1.0, A0 represents the double-layer capacitance Cdl. For PPn-coated electrodes, A0 is the capacitance of serial Cdl and CPPn.Accordingly, the observed impedance changes in ‘PPn-coated’ and‘MI’ sensors can be attributed to the alterations in resistance andcapacitance (Fig. 3b). The hFtn concentration-dependent response

in EIS can be separated into two components: (i) an increase inresistance and (ii) a decrease in capacitance versus the increase inhFtn concentration (Fig. 3c). At the highest hFtn concentration,we observed a 50% increase in resistance and a 20% decrease incapacitance. Non-template BSA did not show the same pattern ofresistance and capacitance changes (Fig. 3c). The change of permit-tivity 1 and resistivity r in the surface materials in response to hFtnbinding (Fig. 3d) is considered the primary mechanism of signalling(re-binding proteins have lower 1 and higher r than the replacedwater in the imprint space, leading to decreased capacitance andincreased resistance).

We also sought to determine if the nanosensor could discrimi-nate between protein conformations. Calmodulin was usedbecause it undergoes measurable conformational changes followingCa2þ binding. Circular dichroism spectroscopy verified that calmo-dulin affinity for Ca2þ was within the micromolar range, as pre-viously reported24–26. Binding of Ca2þ enhanced the a-helicalcontent of calmodulin (Fig. 4a and Supplementary Fig. S4). Thesedata agree with reports demonstrating that the binding of twoions produces the bulk of this large change in the far-UVregion25,26. Our solution binding data are consistent with a Kd of5.6+1.1 mM for two Ca2þ ions binding to calmodulin (Fig. 4b).Interestingly, the Kd is much higher than what appears to be oper-ational in the Ca2þ-bound calmodulin (Ca–CaM) imprint nano-sensor for free Ca2þ (Supplementary Information, Section I.2),where half-saturation is �1 nM (Supplementary Fig. S4). A prob-able explanation is that the off-rate of Ca2þ from the Ca–CaMcomplex bound to the nanotube imprint is much slower than forCa2þ dissociating from the calmodulin complex in solution. Thereis a large excess of calmodulin (0.6 mM) present in the solution

a bGrow nanotube array

Proteinentrapment

Ω

100 nm

Low High

Embed the array

Templateremoval

Low High

Polish the array

Ω

c

Deposit PPn

Carbonsheets

ΩProteinrebinding

Low High

Glass substrate Ti

PPn

hFtn

70 nm

Nanotube

Suppoting polymer

Protein template

PPn coating

50 nm

2 μm

Figure 1 | Fabrication of an MI protein nanosensor. a, Schematic of nanosensor fabrication and template protein detection. The supporting polymer

(SU8-2002) is spin-coated on a glass substrate containing nanotube arrays. Template proteins trapped in the polyphenol (PPn) coating are removed to reveal

the surface imprints. Inset: hypothetical sensor impedance responses at critical stages of fabrication and detection. b, Scanning electron microscopy image

of a polished nanotube array after PPn coating. Inset: cross-section of nanotube tip after polishing. c, TEM images of PPn-coated nanotube tip without (top)

and with (bottom) hFtn.

LETTERS NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.114

NATURE NANOTECHNOLOGY | VOL 5 | AUGUST 2010 | www.nature.com/naturenanotechnology598

© 2010 Macmillan Publishers Limited. All rights reserved.

compared to the available imprints accessible to rebind any releasedCa2þ; however, the motional constraints of the imprint-boundprotein are likely to hinder ion release. That said, Ca–CaM wasdetected by DPV under a variety of free Ca2þ concentrations withthe peak current decreasing as a function of free Ca2þ with a 50%decrease corresponding to subnanomolar free Ca2þ (Fig. 4c).

To further demonstrate the applicability of the nanosensor, wedeveloped a MI using human papillomavirus derived E7 protein(type-16). Analysis revealed the detection of E7 protein at sub-pg l21 levels (Fig. 4d). Importantly, the human papillomavirus E6protein (type-16) was not recognized by the E7 imprint (Fig. 4d).Of note, computational analysis of the interaction between PPnand the template E7 protein revealed several low-energy PPn3-mer dockings on the protein (Supplementary Fig. S5).

In conclusion, we describe a nanosensor that offers selective,label-free, electrochemical detection of proteins by MI. In contrastto conventional MI-based sensors, a key architectural improvementis the non-conductive PPn nanocoating on the nanotube tips. The

level of detection (for example, �10 pg l21 was achieved for hFtn)surpasses that afforded by conventional MI sensors and is compar-able to nanosensors based on biomolecular recognition27,28. Wedemonstrate the potential clinical utility of the nanosensor in theselective recognition of human papillomavirus E7 protein. Thenanosensor design should also prove highly useful in the detectionof other pathogens and toxins, in diagnosing human diseases(through the detection of disease biomarkers), and in a host of pro-teomic applications29.

MethodsNanotube array preparation. Vertically aligned nanotubes were prepared ontitanium-coated glass substrates as described previously14. The nanotube arrays wereembedded in SU8-2002 (SU8) photoresist by spin-coating, then mechanicallypolished to expose the tips. SU8 was applied onto an array at 3,000 r.p.m. for 30 s.Following soft baking at 100 8C (5 min), SU8 was crosslinked by exposure to UV lightfor 3 min and the sample baked at 150 8C overnight. The array was polished with avibratory polisher (Buehler) at a power level of 80% for 6–9 h until the desired patternemerged from the SU8 coating, as monitored by scanning electron microscopy.

200

100

100 300200 400 0.6

35

Concentration (g l−1)

10−13 10−12 10−11 10−10 10−9 10−8 10−7 10−6

hFtnBSA

10−5 10−4 10−3

45

55

65758595

105

60

Peak

cur

rent

(%)

Peak

cur

rent

(%)

BeefBeef+hFtn

hsFtn

hsFtn+hFtn

ahsFtn

ahsFtn+hFtn

70

80

90

100

Ferritin-free

1

10

100 pg l−1

0.4Voltage (V)

0.20.0

−1.5

−1.0

Cur

rent

(μA

)

−0.5

0.0

00

–Zim

ag (kΩ

)

Zreal (kΩ)

150

100

50

250200 150100500

0

–Zim

ag (kΩ

)

Zreal (kΩ)

a b

c

d

Figure 2 | Detection of hFtn using an MI nanosensor. a, Nyquist plots of nanosensor response ( f scanned from 1 Hz to 1 MHz). Traces correspond to

measurements without PPn (red), with hFtn entrapped in PPn (blue), with hFtn removed from PPn (green) and following application of BSA (upper graph,

black line) at concentrations from 1 × 10210 to 1 × 1024 g l21 and hFtn (lower graph, dotted lines) at concentrations from 1 × 10212 to 1 × 1027 g l21. Data

represent three independent determinations. b, DPV current responses to different hFtn concentrations. c, Protein concentration-dependent peak current

response represented as the percentage of peak current (%) versus concentration (g l21) on log–log scales. d, DPV current responses to hFtn in binary

mixtures. The final concentration of hFtn and individual interferent proteins corresponded to 1 × 1028 and 1 × 1026 g l21, respectively. Beef, bovine muscle

protein extract; hsFtn, horse ferritin; ahsFtn, horse apoferritin; hFtn, human ferritin. Data are represented as standard error of the mean (n¼ 3).

NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.114 LETTERS

NATURE NANOTECHNOLOGY | VOL 5 | AUGUST 2010 | www.nature.com/naturenanotechnology 599

© 2010 Macmillan Publishers Limited. All rights reserved.

+−−−

++

+−

+−

−−−

−

− +

+

+

εH2Oρbuffer

εH2O/PPnρbuffer/PPn

εPPnρPPn

εProtein/PPnρProtein/PPn

IZI

IZIIZI

IZI

3.1

CNT-only PPn-coated MI

550

a

c

b

0 2050

70

60

80

90

100250

300

350

400

Referenceelectrode

Workingelectrode

Ru

Rp

Rp Ferritin

Rp BSAA0 BSA

A0 Ferritin

40 60 80 100 120

Concentration (ng l−1)

144.7 178.4

53.9

307.9

Rp (kΩ)

A0 (S.sα0)

R p (k

Ω)

A0

(S.sα0

)

Nanotube

Buffer

PPn

Templateprotein

CPEn = An(jω)−αn

CPE1 : A1, α1

CPE0 : A0, α0

d

Figure 3 | Detection mechanism of the MI nanosensor. a, Equivalent circuit

model with constant phase elements (CPEs). b, Rp and A0 values derived

from Nyquist plot fittings at various fabrication stages: nanotube tips devoid

of PPn (CNT only); nanotube tips coated with template protein entrapped

PPn (PPn-coated); template-protein molecular-imprinted nanotube tips (MI).

c, Rp and A0 versus hFtn or BSA concentrations; error bars indicate standard

error of the mean (n¼ 3). d, Four scenarios of sensor surface conditions

closely related to resistivity (r) and permittivity (1): (i) bare nanotube

surface, where solution double-layer dominates the surface (1H2O, rbuffer);

(ii) nanotube surface coated with PPn, (1PPn, rPPn); (iii) nanotube coated

with PPn with imprint sites (1H2O/PPn, rbuffer/PPn); (iv) imprint sites occupied

with rebound template protein, (1protein/PPn, rprotein/PPn). The differently sized

IZI reflect the relative values of impedance.

190

6,000

5,000

4,000

3,000

2,000

1,000

00 20 40 60

Ca2+ (μm)80 100

10−14 10−13 10−12 10−11

Concentration (g l−1)10−10 10−9 10−8 10−7 10−6

10−11 10−10 10−9 10−8

Free [Ca2+] (M)10−7 10−6 10−5 10−4 10−3 10−2

−10,000

−20,000

10,000

[ε] (

deg

cm2 d

mol

−1)

Δ[ε]

222 (

deg

cm2 d

mol

−1)

0

20,000

30,000

40,000a

b

d

c

200 210 220λ (nm)

230 240 250

110

100

90

80

70E6−16E7−16

60

100

90

80

70

60

50

40

Peak

cur

rent

(%)

Peak

cur

rent

(%)

Imprint independent current

Without CaM (EGTA only)

Ca2+

Figure 4 | Detection of calcium-dependent calmodulin conformational

changes and HPV-derived oncoproteins. a, Far-UV circular dichroism

spectra showing the molar ellipticity (measured in deg cm2 dmol21) for

calmodulin in the absence (black) and presence of 0.9 (blue), 2.7 (purple),

5.3 (green), 13.9 (grey) and 30.4 (red) mM Ca2þ. The arrow indicates

increasing negative ellipticity at 222 nm. b, Change in calmodulin mean

residue ellipticity at 222 nm ([1]0–[1]Ca2þ) as a function of added Ca2þ,

fitted with a sigmoidal curve. The increasing value for D[1] indicates an

increase in helix as Ca2þ is added. c, Nanosensor DPV represented as the

percentage of peak current versus free Ca2þ concentration. The peak current

is normalized to that recorded in Tris buffer with 1 mM EGTA (see

Supplementary Information for details). The imprint-independent current

represents the total residual current with overloaded Ca2þ. CaM, calmodulin.

Inset: conformational state of CaM following Ca2þ binding. d, MI recognition

of oncoprotein E7 type-16 by an E7-16 imprinted sensor. Protein E6 type-16

was used as a negative control. Data represent percentage of peak current

versus E7 (open diamonds) or E6 (closed circles) type-16 protein

concentrations. Error bars indicate standard error of the mean (n¼ 3).

LETTERS NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.114

NATURE NANOTECHNOLOGY | VOL 5 | AUGUST 2010 | www.nature.com/naturenanotechnology600

© 2010 Macmillan Publishers Limited. All rights reserved.

Electropolymerization of PPn onto nanotube array tips. PPn film was depositedon the exposed nanotube tips by cyclic voltammetry in phosphate buffered saline(PBS) containing 1.5 mM phenol (pH¼ 7.4). To the working electrode (nanotubetips) a ‘ramping’ voltage was applied at a scanning rate of+50 mV s21 between 0.0to 0.9 V versus the reference electrode (Ag/AgCl) for five cycles. An additionalelectrode consisting of a platinum wire was connected as the counter electrode forcurrent injection. To entrap template proteins in the PPn coating, 100 mg ml21 ofthe corresponding template protein was added to the PPn deposition buffer. A300 mV d.c. voltage (30 s) was first used to attract proteins onto the nanotube tips,followed by application of five cycles of voltage scanning, as described above.Note that, during entrapment of calmodulin in the PPn coating, 1 mM Ca2þ wasincluded to render conformation with a full-scale elongation (‘open’) that offered adistinctive imprint morphology by its globular shape at Ca2þ free or other partial‘close’ status.

Electrochemical measurements. A three-electrode electrochemical system wasconfigured by connecting the titanium film (on which nanotubes were verticallygrown) as the working electrode, using Ag/AgCl as the reference electrode, and withthe platinum wire serving as the counter electrode. EIS was used to monitorimpedance changes of the electrode surface and its interface to the buffer solutioncontaining 1 mM ferrocyne carboxyl acid in PBS during PPn deposition. An a.c.sine-wave voltage was applied to the chip with a superimposed d.c. voltage at300 mV. Its peak-to-peak amplitude was 10 mV. The frequency varied from 1 Hz to1 MHz with 20 point impedance readings per decade. The impedance data were thenrepresented in the form of Nyquist plots, which were fitted to an equivalent circuitby the analysis function in Echem Analyst software. DPV was conducted in bindingbuffer supplemented with 1 mM ferrocyne carboxyl acid. The initial and finalpotentials versus the reference electrode were 0.0 and 0.5 V, respectively. The pulsesize was 50 mV, and the pulse time 0.05 s; the step size was 2 mV and the sampleperiod 0.1 s. For imprint refilling experiments, the nanotube arrays were coated withPPn only or with PPn plus hFtn. A leakage current was observed for the PPn-onlysample after imprint development. By eliminating the charge due to this leakagecurrent in the PPn-only sample, the imprint-contributed charge in the imprintedsample was obtained, to be used in estimating the number of imprints on the chip.

Protein imprint development. For imprint development, the nanotube tip arraywith protein-entrapped PPn coating was rinsed and incubated overnight indeionized water at room temperature. In some instances, a developing buffercontaining 5% acetic acid and 10% sodium dodecyl sulphate was used, whichresulted in higher protein extraction efficiency. After protein entrapment andremoval, the sample was evaluated by TEM and EIS. Before measuring template-protein rebinding, DPV was repetitively performed in a template-protein-free buffersolution until the current stabilized. Nanosensor responses were then measured byDPV or EIS following incubation for 30 min at room temperature with the templateprotein at various concentrations.

Received 9 March 2010; accepted 14 May 2010;published online 27 June 2010

References1. Wulff, G. & Sharhan, A. Uber die Anwendung von enzymanalog gebauten

Polymeren zur Racemattrennung. Angew. Chem. 84, 364 (1972).2. Vlatakis, G., Anderson, L. I., Muller, R. & Mosbach, K. Drug assay using

antibody mimics made by molecular imprinting. Nature 361, 645–647 (1993).3. Hoshino, Y., Kodama, T., Okahata, Y. & Shea, K. J. Peptide imprinted polymer

nanoparticles: a plastic antibody. J. Am. Chem. Soc. 130, 15242–15243 (2008).4. Bossi, A., Bonini, F., Turner, A. P. F. & Piletsky, S. A. Molecularly imprinted

polymers for the recognition of proteins: the state of the art. Biosens. Bioelectron.22, 1131–1137 (2007).

5. Karube, J. et al. Molecular recognition in continuous polymer rods prepared by amolecular imprinting technique. Anal. Chem. 65, 2223–2224 (1993).

6. Kempe, M. Antibody-mimicking polymers as chiral stationary phases in HPLC.Anal. Chem. 68, 1948–1953 (1996).

7. Rashid, B. A., Briggs, R. J., Hay, J. N. & Stevenson, D. Preliminary evaluation of amolecular imprinted polymer for solid-phase extraction of tamoxifen. Anal.Commun. 34, 303–305 (1997).

8. Shi, H., Tsai, W.-B., Garrison, M. D., Ferrari, S. & Ratner, B. D. Template-imprintednanostructured surfaces for protein recognition. Nature 398, 593–597 (1999).

9. Rick, J. & Chou, T. C. Using protein templates to direct the formation of thin-film polymer surfaces. Biosens. Bioelectron. 22, 544–549 (2006).

10. Wang, Y. et al. A potentiometric protein sensor built with surface molecularimprinting method. Biosens. Bioelectron. 24, 162–166 (2008).

11. Bossi, A., Piletsky, S. A., Piletska, E. V., Righetti, P. G. & Turner, A. P. F. Surface-grafted molecularly imprinted polymers for protein recognition. Anal. Chem. 73,5281–5286 (2001).

12. Zheng, G., Patolsky, F., Cui, Y., Wang, W. U. & Lieber, C. M. Multiplexedelectrical detection of cancer markers with nanowire sensor arrays. NatureBiotechnol. 23, 1294–1301 (2005).

13. Yu, X. et al. Carbon nanotube amplification strategies for highly sensitiveimmunodetection of cancer biomarkers. J. Am. Chem. Soc. 128,11199–11205 (2006).

14. Yu, Y. et al. Assembly of multi-functional nanocomponents on periodicnanotube array for biosensors. Micro. Nano. Lett. 4, 27–33 (2009).

15. Riskin, M., Tel-Vered, R. & Willner, I. The imprint of electropolymerizedpolyphenol films on electrodes by donor–acceptor interactions: selectiveelectrochemical sensing of N,N ′-dimethyl-4,4′-bipyridinium (methyl viologen).Adv. Func. Mater. 17, 3858–3863 (2007).

16. Lakshmi, D. et al. Electrochemical sensor for catechol and dopamine based on acatalytic molecularly imprinted polymer-conducting polymer hybridrecognition element. Anal. Chem. 81, 3576–3584 (2009).

17. Panasyuk, T. L., Mirsky, V. M., Piletsky, S. A. & Wolfbeis, O. S.Electropolymerized molecularly imprinted polymers as receptor layers incapacitive chemical sensors. Anal. Chem. 71, 4609–4613 (1999).

18. Choong, C. L., Bendallb, J. S. & Milnea, W. I. Carbon nanotube array: a new MIPplatform. Biosen. Bioelectron. 25, 652–656 (2009).

19. Wei, Y., Qiu, L., Owen, C. & Lai, E. P. C. Encapsulation of quantum dots andcarbon nanotubes with polypyrrole in a syringe needle for automatedmolecularly imprinted solid phase pre-concentration of ochratoxin A in redwine analysis. Sens. Instrum. Food Qual. 1, 133–141 (2007).

20. Xie, C., Li, H., Li, S., Wu, J. & Zhang, Z. Surface molecular self-assemblyfor organophosphate pesticide imprinting in electropolymerizedpoly(p-aminothiophenol) membranes on a gold nanoparticle modified glassycarbon electrode. Anal. Chem. 82, 241–249 (2010).

21. Tabard-Cossa, V., Trivedi, V., Wiggin, M., Jetha, N. N. & Marziali, A. Noiseanalysis and reduction in solid-state nanopores. Nanotechnology 18,305505 (2007).

22. Gong, K., Chakrabarti, S. & Dai, L. Electrochemistry at carbon nanotubeelectrodes: is the nanotube tip more active than the side wall? Angew. Chem. Int.Ed. 47, 5446–5450 (2008).

23. Bartlett, P. N., Tebbutt, P. & Tyrrell, C. H. Electrochemical immobilization ofenzymes. 3. Immobilization of glucose oxidase in thin films of electrochemicallypolymerized phenols. Anal. Chem. 64, 138–142 (1992).

24. Klee, C. B. Conformational transition accompanying the binding of Ca2þ to theprotein activator of 3′ ,5′-cyclic adenosine monophosphate. Biochemistry 16,1017–1024 (1977).

25. Crouch, T. H. & Klee, C. B. Positive cooperative binding of calcium to bovinebrain calmodulin. Biochemistry 19, 3692–3698 (1980).

26. Weinstein, H. & Mehler, E. L. Ca2þ-binding and structural dynamics in thefunctions of calmodulins. Annu. Rev. Physiol. 56, 213–236 (1994).

27. Devanathan, S. et al. Subpicomolar sensing of d-opioid receptor ligands bymolecular-imprinted polymers using plasmon-waveguide resonancespectroscopy. Anal. Chem. 77, 2569–2574 (2005).

28. Patolsky, F., Zheng, G. & Lieber, C. M. Fabrication of silicon nanowire devicesfor ultrasensitive, label-free, real-time detection of biological and chemicalspecies. Nature Protoc. 1, 1711–1724 (2006).

29. Sanchez-Carbayo, M. Antibody arrays: technical considerations and clinicalapplications in cancer. Clin. Chem. 52, 1651–1659 (2006).

AcknowledgementsThe authors acknowledge financial support from The Seaver Institute (T.C.C and M.J.N.).L.R. was supported by the China Scholarship Council (no. 2008677005). We thank O.Gursky and S. Jayaraman (Boston University School of Medicine, Department ofPhysiology and Biophysics) for use of their Aviv spectropolarimeter.

Author contributionsD.C. contributed to the original MI nanosensor concept, overall experimental design, dataanalysis, manuscript preparation and directing of the measurements. L.R. was responsiblefor the EIS recordings and contributed to sensor fabrication. H.Z. fabricated the nanotubearrays. C.X. was responsible for protein preparation and purification. Y.Y. contributed todevelopment of the PPn nanocoating. H.W. and Y.L. helped with the high-resolution TEMimage. M.F.R. was responsible for the circular dichroism measurements. L.Z. and J.H.C.were responsible for computational analysis of the interaction between the E7 protein andPPn. M.J.N. provided technical support for nanotube fabrication and assisted inmanuscript editing. Z.R. contributed expertise regarding nanotube fabrication,experimental design for TEM evaluation of hFtn entrapment, and editing of themanuscript. T.C.C. contributed to the design of experiments for demonstrating nanosensorselectivity and was responsible for writing and editing the revised manuscript.

Additional informationThe authors declare no competing financial interests. Supplementary informationaccompanies this paper at www.nature.com/naturenanotechnology. Reprints andpermission information is available online at http://npg.nature.com/reprintsandpermissions/.Correspondence and requests for materials should be addressed to D.C.

NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.114 LETTERS

NATURE NANOTECHNOLOGY | VOL 5 | AUGUST 2010 | www.nature.com/naturenanotechnology 601

© 2010 Macmillan Publishers Limited. All rights reserved.