S

D

Xa

b

c

ARRAA

KHCA

lhe(t2e

Hrsdccct

ju

eH

(

0d

Journal of Virological Methods 171 (2011) 405–407

Contents lists available at ScienceDirect

Journal of Virological Methods

journa l homepage: www.e lsev ier .com/ locate / jv i romet

hort communication

etection of human adenovirus hexon antigen using carbon nanotube sensors

ue-Hai Jina,∗, Atsushi Ishiib, Koki Aokia, Susumu Ishidaa, Koichi Mukasab, Shigeaki Ohnoc

Department of Ophthalmology, Hokkaido University Graduate School of Medicine, Kita 15, Nishi 7, Kita-ku, Sapporo 060-8638, JapanCreative Research Initiative “Sousei”, Hokkaido University, Kita 21, Nishi 10, Kita-ku, Sapporo 001-0021, JapanDepartment of Ocular Inflammation and Immunology, Hokkaido University Graduate School of Medicine, Kita 15, Nishi 7, Kita-ku, Sapporo 060-8638, Japan

rticle history:eceived 21 March 2010eceived in revised form 5 December 2010ccepted 13 December 2010vailable online 17 December 2010

a b s t r a c t

Human adenoviruses (HAdVs) have been implicated in a wide range of diseases affecting primarily therespiratory, ocular and gastrointestinal systems. A rapid and efficient method for the detection of HAdVhexon antigen is described using carbon nanotube (CNT) sensors. Anti-HAdV antibody was immobilisedon the reverse surface of a CNT sensor. As a control, non-specific mouse IgG was immobilised on another

eywords:uman adenovirusesarbon nanotube sensorntigen–antibody interaction

CNT sensor. I–Vgate curves were measured after incubation of various concentrations of recombinantHAdVs hexon antigen with anti-HAdVs antibody-immobilised or non-specific mouse IgG-immobilisedsensors. The curves showed a positive shift that was dependent on the hexon antigen concentration in theanti-HAdV antibody-immobilised sensor, whereas no such shift was observed in the non-specific mouseIgG-immobilised sensor. The sensitivity of the CNT sensor method was greater than that of enzyme-linked immunosorbent assay. Hence, this method offers a new tool for HAdV detection by analysing

tions

antigen–antibody interac

Human adenoviruses (HAdVs) circulate widely and infect mil-ions of people worldwide. HAdVs have been implicated in variousuman infections including gastroenteritis, respiratory infections,ye infections, acute hemorrhagic cystitis and meningoencephalitisMena and Gerba, 2009). The HAdV family consists of 52 serotypeshat are grouped into 6 species (McCarthy et al., 2009; Moro et al.,009). Several new serotypes have been identified recently (Aokit al., 2008; Ishiko et al., 2008).

Adenoviral conjunctivitis is caused primarily by HAdV-3,AdV-4, HAdV-8, HAdV-19 and HAdV-37 serotypes. Adenovi-

al conjunctivitis is diagnosed clinically on the basis of history,ymptoms and signs. However, conjunctivitis caused by Chlamy-ia and herpes simplex virus, allergic conjunctivitis and bacterialonjunctivitis present symptoms similar to those of adenoviralonjunctivitis. Therefore, it is impossible to distinguish adenoviralonjunctivitis from other types of conjunctivitis by clinical symp-

oms alone (Sambursky et al., 2006; Kaneko et al., 2008).

The “gold standard” for laboratory diagnosis of adenoviral con-unctivitis is isolation of the virus in cell cultures. However, thissually requires 1–2 weeks for completion. Polymerase chain

Abbreviations: HAdVs, human adenoviruses; CNT, carbon nanotube; ELISA,nzyme-linked immunosorbent assay; Isd, the current between the source and drain;RPO, horse radish peroxidase.∗ Corresponding author. Tel.: +81 11 716 1161; fax: +81 11 706 5948.

E-mail addresses: sapporo031005@yahoo.co.jp, drjin2001@yahoo.com.cnX.-H. Jin).

166-0934/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.jviromet.2010.12.004

.© 2010 Elsevier B.V. All rights reserved.

reaction (PCR) assays provide diagnostic tests for adenoviral con-junctivitis that are more rapid and sensitive than cell culture(Percivalle et al., 2003). However, these laboratory tests are per-formed rarely for conjunctivitis because of the considerable timerequired for obtaining test results and the self-limiting nature ofthe disease (Sambursky et al., 2006). Immunochromatography hasbeen regarded as a simple and rapid test to detect HAdV anti-gen with high specificity; however, its sensitivity is only 60–70%(Kaneko et al., 2008). Some other rapid tests such as enzymeimmunoassays, immunofiltration, direct immunofluorescence andimmune electron microscopy are not employed commonly becauseof their lack of sensitivity and specificity, as well as because oftheir technical complexity (Percivalle et al., 2003; Sambursky et al.,2006).

Biosensors have been employed to detect biomolecules suchas proteins and nucleotides, micro-organisms as well as wholebiological tissues (Balasubramanian and Burghard, 2006; Gruner,2006). A biosensor incorporates a biocomponent coupled to atransducer, which translates the interaction between the ana-lyte and the biocomponent into a signal that can be processedand detected. A wide range of transducers has been employedin biosensors of which electrochemical and optical transducersare the most common (Reardon et al., 2009). The recent boom

in nanofabrication technology and biofunctionalisation methodsfor carbon nanotubes (CNTs) has encouraged significant researchinterest in the development of CNT-based biosensors for monitor-ing biorecognition events and biocatalytic processes. The uniqueproperties of CNTs, which are rolled-up sheets of carbon atoms with

406 X.-H. Jin et al. / Journal of Virological Methods 171 (2011) 405–407

Fw

aitoaRbbTtbo

adpwot(

oahgcorepmn

totistappta

tsf14

Fig. 2. Detection of binding of HAdV hexon antigen to the immobilised anti-HAdV antibody using CNT sensors. (a) Effects of binding of the hexon antigento immobilised anti-HAdV antibody on I–Vgate curves. (b) Effects of binding ofthe hexon antigen to immobilised non-specific mouse IgG on I–Vgate curves. Thenumbers within the frames of (a) and (b) represent log10[recombinant hexon con-centration (mg/ml)]. (c) Current (Isd) is plotted at 5 V of I–Vgate curves (a) and 2 Vof I–Vgate curves (b). The closed and open circles represent data obtained from

ig. 1. Schematic model of I–Vgate measurement using CNT sensors. I–Vgate curvesere measured in air and a voltage of −20 V to +20 V was applied to the back gate.

diameter of approximately 1 nm, offer excellent prospects fornterfacing biorecognition events with electronic signal transduc-ion. CNT-based biosensors are extremely sensitive and can detectnly a few or even a single molecule of a chemical or biologicalgent (Luong et al., 2007; Balasubramanian and Burghard, 2006;ivas et al., 2007). Takeda et al. (2005) developed a CNT-basediosensor to detect the binding of hemagglutinins, the mem-rane proteins of influenza virus, to anti-hemagglutinin antibodies.hey demonstrated that the CNT device had higher sensitivityhan enzyme-linked immunosorbent assay (ELISA). The interactionetween hemagglutinins and anti-hemagglutinin antibodies wasbserved by measuring I–Vgate curves.

In this study, a CNT-based biosensor to detect HAdV hexonntigen by immobilising anti-HAdV antibody on the sensor waseveloped and the sensitivity of the CNT sensor method was com-ared with that of ELISA. The CNT field-effect transistor deviceas manufactured by a combination of position-controlled growth

f CNTs, electron-beam lithography and electron-beam deposi-ion. The details of this procedure have been described previouslyTakeda et al., 2005).

For this study, HAdV-19 was isolated from a conjunctival swabf a patient with acute conjunctivitis. The viral DNA was extractedfter propagation of the isolate in HeLa cells and A549 cells. Theexon gene of HAdV-19a was amplified using PCR. The amplifiedene was cloned into a baculovirus expression vector. Next, thisonstruct was transfected into Tn5 insect cells for the expressionf the recombinant hexon antigen. The culture media were thenecovered and the cytoplasmic lysates of the transfected cells werextracted. Anti-HAdV monoclonal antibody in mouse ascites wasurified by protein-A sepharose 4B (Sigma–Aldrich) affinity chro-atography. A non-specific mouse IgG was also prepared as the

egative control.In order to examine the properties of CNTs, the current between

he source and drain (Isd) was measured in air by applying a voltagef −20 V to +20 V to the back gate that faced the reverse surface ofhe CNTs. The schematic model of I–Vgate measurements is shownn Fig. 1. Isd was plotted as a function of the potential between theource and drain (data not shown). An Agilent 4155C semiconduc-or parameter analyser was employed to measure I–Vgate curvest room temperature. CNTs exhibit conductive or semi-conductiveroperties depending on their chirality. Isd is dependent on theotential of the back gate. In this experiment, Isd increased whenhe back gate potential decreased to −20 V, suggesting that Isd wasffected strongly by the effective potential surrounding the CNT.

Before modification of the reverse surface of the CNT sensor,he surface was treated with 2 N NaOH to obtain an oxidized

urface. 3-Mercaptopropyltriethoxysilane was added to the sur-ace at 45 ◦C for 15 min and then heated to 200 ◦C for 30 min.mM sodium tetrahydroborate was then added to the surface at5 ◦C for 15 min. After washing with deionised water, the pre-

I–Vgate curves (a) and (b), respectively. The closed triangle represents absorbancevalues measured at 450 nm after terminating the HRPO reaction in the ELISAmethod.

ogical

i4tNsfustt1waa

tattapsIgIIgHcccotTtas

cTtms

sAInwFttl

antibody on a carbon nanotube sensor. J. Nanosci. Nanotechnol. 7, 752–

X.-H. Jin et al. / Journal of Virol

ncubated mixture of anti-HAdV antibody and sulfosuccinimidyl-(N-maleimidomethyl) cyclohexane-1-carboxylate was added tohe surface and allowed to stand at room temperature for 30 min.ext, the anti-HAdV antibody was immobilised on the CNT sensor

urface. The non-specific mouse IgG was immobilised on the sur-ace of another CNT sensor in the same way and the sensor wassed as a negative control. After washing with phosphate-bufferedaline (PBS) buffer and deionised water, 50 �l of various concen-rations of recombinant hexon antigen ranging from 1.0 × 10−11

o 1.0 × 10−5 mg/ml were added to the surface of the 2 sensors for5 min at 25 ◦C. The surface of the sensor was washed with PBS, andas dried using nitrogen gas. Next, I–Vgate curves were measured

t each step of the reaction between hexon antigen and anti-HAdVntibody. The measurement required about 1 min for a result.

Increasing the concentration of hexon antigen solution addedo the sensor resulted in a positive shift of I–Vgate curves of thenti-HAdV antibody-immobilised CNT sensor (Fig. 2a), suggestinghat the effective potential surrounding of the CNT may changeo positive due to the formation of a complex between the hexonntigen and the anti-HAdV antibody. The same experiment waserformed using 3 separate CNT sensors and a similar positivehift of I–Vgate curves was observed. For the control experiment,–Vgate curves were plotted after addition of series of hexon anti-en solutions of various concentrations to the non-specific mousegG-immobilised surface (Fig. 2b). There was no apparent shift in–Vgate curves or increase in Isd after the addition of the hexon anti-en to the non-specific mouse IgG-immobilised surface (Fig. 2b).ence, unlike the correlation indicated in Fig. 2a, there was no clearorrelation between Isd and the concentration of hexon antigen. Tolarify the shift of I–Vgate curves, we plotted Isd at 5 V of Vgate. I–Vgate

urves shifted in the positive direction in the concentration rangef 1.0 × 10−11 to 1.0 × 10−6 mg/ml and in the negative direction inhe concentration range of 1.0 × 10−6 to 1.0 × 10−5 mg/ml (Fig. 2c).his result indicated that the shift of I–Vgate curves might be dueo specific binding of hexon antigen to the immobilised anti-HAdVntibody, and that the binding may change the effective potentialurrounding the CNT to positive.

The negative shift of I–Vgate curves in case of high antigen con-entration has been observed previously (Takeda et al., 2007).he mechanism of this negative shift remains unknown. Hence,o determine the antigen concentration of an unknown sample,

easuring the shift in I–Vgate curves of the sample and its dilutedolutions may be one way to avoid an erroneous result.

Next, ELISA plates were coated with the diluted hexon antigenamples and then incubated the plates with anti-HAdV antibody.fter rinsing with PBS, the plate was incubated with anti-mouse

gG conjugated with horseradish peroxidase (HRPO). After termi-ating the HRPO reaction, the absorbance values of diluted samples

ere measured at 450 nm. The line drawn across the closed circle in

ig. 2c is the tentative calibration line of the CNT sensor. The detec-ion limit of the hexon antigen using the CNT sensor was estimatedo be approximately 10−11 mg/ml on the basis of the calibrationine. Various concentrations of hexon antigen solution were exam-

Methods 171 (2011) 405–407 407

ined using ELISA. The absorbance of the various solutions at 450 nmis shown as closed triangles in Fig. 2c. The line drawn across theclosed triangle is the tentative calibration line of the ELISA. Thedetection limit of the hexon antigen using ELISA was estimated tobe 10−6 mg/ml. Therefore, the sensitivity of the CNT sensor methodfor detecting the hexon antigen was almost 105 times higher thanthat of the ELISA method.

The CNT sensor method reported here is a simple and quickdetection method requiring a total of 16–17 min, including 15 minof reaction time and 1–2 min for drying and measurement. How-ever, the instrument used in the study is expensive and is awkwardto be used in a clinical trial. It is necessary to develop a simple andportable instrument so as to make the CNT sensor method viablein a clinical setting for the detection of HAdV hexon antigen.

References

Aoki, K., Ishiko, H., Konno, T., Shimada, Y., Hayashi, A., Kaneko, H., Ohguchi, T.,Tagawa, Y., Ohno, S., Yamazaki, S., 2008. Epidemic keratoconjunctivitis due tothe novel hexon-chimeric-intermediate 22,37/H8 human adenovirus. J. Clin.Microbiol. 46, 3259–3269.

Balasubramanian, K., Burghard, M., 2006. Biosensors based on carbon nanotubes.Anal. Bioanal. Chem. 385, 452–468.

Gruner, G., 2006. Carbon nanotube transistors for biosensing applications. Anal.Bioanal. Chem. 384, 322–335.

Ishiko, H., Shimada, Y., Konno, T., Hayashi, A., Ohguchi, T., Tagawa, Y., Aoki, K., Ohno,S., Yamazaki, S., 2008. Novel human adenovirus causing nosocomial epidemickeratoconjunctivitis. J. Clin. Microbiol. 46, 2002–2008.

Kaneko, H., Maruko, I., Iida, T., Ohguchi, T., Aoki, K., Ohno, S., Suzutani, T., 2008.The possibility of human adenovirus detection from the conjunctiva in asymp-tomatic cases during nosocomial infection. Cornea 27, 527–530.

Luong, J.H., Male, K.B., Hrapovic, S., 2007. Carbon nanotube-based electrochemi-cal biosensing platforms: fundamentals, applications, and future possibilities.Recent Pat. Biotechnol. 1, 181–191.

McCarthy, T., Lebeck, M.G., Capuano, A.W., Schnurr, D.P., Gray, G.C., 2009. Moleculartyping of clinical adenovirus specimens by an algorithm which permits detectionof adenovirus coinfections and intermediate adenovirus strains. J. Clin. Virol..

Mena, K.D., Gerba, C.P., 2009. Waterborne adenovirus. Rev. Environ. Contam. Toxicol.198, 133–167.

Moro, M.R., Bonville, C.A., Suryadevara, M., Cummings, E., Faddoul, D., Kobayaa, H.,Branigan, P.J., Domachowske, J.B., 2009. Clinical features, adenovirus types, andlocal production of inflammatory mediators in adenovirus infections. Pediatr.Infect. Dis. J. 28, 376–380.

Percivalle, E., Sarasini, A., Torsellini, M., Bruschi, L., Antoniazzi, E., Grazia Revello,M., Gerna, G., 2003. A comparison of methods for detecting adenovirus type 8keratoconjunctivitis during a nosocomial outbreak in a Neonatal Intensive CareUnit. J. Clin. Virol. 28, 257–264.

Reardon, K.F., Zhong, Z., Lear, K.L., 2009. Environmental applications ofphotoluminescence-based biosensors. Adv. Biochem. Eng. Biotechnol..

Rivas, G.A., Rubianes, M.D., Rodriguez, M.C., Ferreyra, N.F., Luque, G.L., Pedano, M.L.,Miscoria, S.A., Parrado, C., 2007. Carbon nanotubes for electrochemical biosens-ing. Talanta 74, 291–307.

Sambursky, R., Tauber, S., Schirra, F., Kozich, K., Davidson, R., Cohen, E.J., 2006. TheRPS adeno detector for diagnosing adenoviral conjunctivitis. Ophthalmology113, 1758–1764.

Takeda, S., Ozaki, H., Hattori, S., Ishii, A., Kida, H., Mukasa, K., 2007. Detection ofinfluenza virus hemagglutinin with randomly immobilized anti-hemagglutinin

756.Takeda, S., Sbagyo, A., Sakoda, Y., Ishii, A., Sawamura, M., Sueoka, K., Kida, H., Mukasa,

K., Matsumoto, K., 2005. Application of carbon nanotubes for detecting anti-hemagglutinins based on antigen–antibody interaction. Biosens. Bioelectron 21,201–205.