Detection of single-base DNA mutations byenzyme-amplified electronic transduction

Fernando Patolsky, Amir Lichtenstein, and Itamar Willner*

Here we describe a method for the sensitive detection of a single-base mutation in DNA. We assembled aprimer thiolated oligonucleotide, complementary to the target DNA as far as one base before the mutation site,on an electrode or a gold–quartz piezoelectric crystal. After hybridizing the target DNA, normal or mutant, withthe sensing oligonucleotide, the resulting assembly is reacted with the biotinylated nucleotide, complementaryto the mutation site, in the presence of polymerase. The labeled nucleotide is coupled only to the double-stranded assembly that includes the mutant site. Subsequent binding of avidin–alkaline phosphatase to theassembly, and the biocatalyzed precipitation of an insoluble product on the transducer, provides a means toconfirm and amplify detection of the mutant. Faradaic impedance spectroscopy and microgravimetric quartz-crystal microbalance analyses were employed for electronic detection of single-base mutants. The lower limitof sensitivity for the detection of the mutant DNA is 1 × 10-14 mol/ml. We applied the method for the analysis ofpolymorphic blood samples that include the Tay–Sachs genetic disorder.The sensitivity of the method enablesthe quantitative analysis of the mutant with no PCR pre-amplification.

RESEARCH ARTICLE

http://biotech.nature.com • MARCH 2001 • VOLUME 19 • nature biotechnology

Sensitivity and selectivity are two important measures in the evalua-tion of DNA biosensor devices1. The development of bioelectronicDNA analysis systems has attracted substantial research efforts in theareas of DNA biosensors2–4 applied to gene analysis, detection ofgenetic disorders, tissue matching, and forensic applications, andDNA computers5–7 for the selective detection of specific basesequences. Optical8, electrochemical9–13, and microgravimetric,quartz-crystal microbalance14 transduction methods have beenreported for detection of oligonucleotide–DNA hybridization. Partialselectivity in DNA sequence detection was reported by the applicationof an oligonucleotide probe capable of forming a single double-stranded turn, and exhibiting four to six base mismatches with respectto the analyte DNA. The substantially different stabilities of the result-ing complementary strands allowed the detection of the respectivemismatches12,14. Recently, an electrochemical method to sense a single-base mutation in an oligonucleotide was demonstrated, using anoligonucleotide-tagged redox enzyme and an oligonucleotide-associ-ated redox polymer as a sensing and electrical contacting matrix15. Theeffectiveness of electrical contacting, and the resulting amperometricresponse were found to be controlled by the number of base mis-matches between the probe oligonucleotide and the analyte gene. Theelectronic transduction of a single-base mutation remains, however, afundamental issue in bioelectronics. Mismatch mutations occurringupon transition or transversion of nucleotide bases often result infunctional consequences in the protein expressed by the gene mutant.Similarly, point mutations of proto-oncogenes may lead to malignanttransformations of the cell. For example, it is established that inhuman bladder carcinoma, a G to T transversion in codon 12 of theHa-ras and Ki-ras proto-oncogene converts them into oncogenes16–18.

Recently, we reported that biocatalyzed precipitation on electronictransducers (e.g., electrodes or piezoelectric crystals) provides ameans to amplify biochemical detection events. Enzymes or oligonu-cleotides attached to electrodes or quartz crystals were found to act assensitive sensing interfaces for the respective substrates19, or comple-mentary DNA20. Here we report a method for the enzyme-amplifiedelectronic transduction of a single-base mutation in DNA. We applyFaradaic impedance spectroscopy and microgravimetric quartz-crys-tal microbalance analyses, to detect the DNA mutant. Faradaic

impedance spectroscopy is a useful electrochemical technique to fol-low the capacitance and electron transfer resistance properties of theelectrode/electrolyte interface. These properties are elucidated by theapplication of an alternate potential on the sensing electrode and theelucidation of the capacitance and electron transfer resistance prop-erties of the electrode interface (see below). Microgravimetric quartz-crystal-microbalance analyses utilize a piezoelectric quartz crystal asthe electronic transducer. The resonance frequency of the crystal iscontrolled by its mass, and any mass changes occurring on the crystalas a result of biorecognition events or the biocatalyzed precipitationof an insoluble product are reflected by changes in the resonance fre-quency of the crystal. The electronic transduction of the single-basemutation in the DNA proceeds with sensitivity, and enables the quan-titative detection of the mutant.

Results and discussionA scheme describing the analysis of a single-base mutation in anoligonucleotide is shown in Figure 1. The detection of the 41-basemutant oligonucleotide (1), which includes a mutation, the substitu-tion of a guanine for an adenine in the normal sequence (2), is illus-trated. The thiolated oligonucleotide (3), which is complementary tothe oligonucleotide fragment of (1) or (2) as far as the point of muta-tion, is used as the oligonucleotide probe. The assembly of the probe(3) onto the transducer (e.g., gold electrode or gold-quartz crystal)yields the sensing interface. Interaction of the sensing interface withthe mutant (1) or the oligonucleotide representing the normalsequence (2) generates the respective double-stranded assembly onthe transducer. The resulting interface is then reacted with thebiotinylated base complementary to the mutant (e.g., biotinylatedcytosine triphosphate, b-dCTP) in the presence of DNA polymerase(Klenow fragment). (The Klenow fragment exhibits exonucleaseactivity for 3'–5' single-stranded DNA. Because the 5′ position of theprobe is linked to the electrode, the biocatalyst reveals only poly-merase activity.) In the presence of the double-stranded assemblythat includes the mutant, surface coupling of b-dCTP to the probeoligonucleotide proceeds. The resulting assembly is then interactedwith an enzyme-linked avidin conjugate, at which point the enzymecatalyzes the precipitation of an insoluble product on the electrode

Institute of Chemistry, The Hebrew University of Jerusalem, Jerusalem 91904, Israel. *Corresponding author (willnea@vms.huji.ac.il).

253

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://b

iote

ch.n

atu

re.c

om

© 2001 Nature Publishing Group http://biotech.nature.com

support. In the present study, avidin–alkaline phosphatase (A-Alk.Ph) is used as the conjugate, and its association with the detec-tor interface catalyzes the oxidative hydrolysis of 5-bromo-4-chloro-3-indoyl phosphate (4), to the insoluble indigo derivative (5).Precipitation occurs only if the single-base mutant hybridizes to theprobe linked to the sensing interface. The biocatalyzed precipitationof (5) provides a means to amplify the detection process, and theextent of precipitate formed on the transducer is controlled by theamount of DNA mutant associated with the sensing interface, and thetime interval employed for the biocatalyzed precipitation of (5). Theinsoluble product formed on the electrode alters the capacitance andelectron transfer resistance at the conductive support. Similarly, thebiocatalyzed precipitation of (5) on the gold-quartz crystal can bedetected microgravimetrically by following the frequency changes ofthe piezoelectric crystal.

The probe oligonucleotide (3) was assembled on the gold electrode(or the gold-quartz crystal) to yield a surface coverage of∼ 2.3 × 10-11 mol/cm2. The surface coverage at time intervals of modi-fication of the gold surfaces was determined by chronocoulometry21

or by quartz-crystal microbalance measurements. One method toassess the formation of the oligonucleotide monolayer, the hybridiza-tion of the analyte to the probe DNA, the association of theavidin–alkaline phosphatase conjugate with the b-dCTP, and the precipitation of the insoluble producton the electrode, is Faradaic impedance spec-troscopy. Using this method, an alternating potentialof low amplitude and variable frequency is appliedto the electrode, close to the redox potential of theredox label in solution. The imaginary impedance(Zim) and the real impedance (Zre) are recorded as afunction of the applied frequency with an electro-chemical impedance analyzer. The resulting Nyquistplot (Zim vs. Zre) consists of a semicircular region,characteristic of the low-frequency domain,followed by a linear portion characteristic of thehigh-frequency region of the diffusion-controlledelectron transfer. The diameter of the semicircle (x-axis) corresponds to the electron transfer resistanceat the electrode. Thus it is anticipated that chemicalmodification of the electrode by the immobilizationof DNA or proteins, or the precipitation of an insol-uble product, will change the electron transfer resis-tance of the electrode interface.

Figure 2A (curve a) shows the Faradaic impedance spectra, usingFe(CN)6

3-/Fe(CN)64- as redox label, of the sensing interface. Curve

(b) shows the Faradaic impedance spectra of the sensing interfaceafter interaction with the mutant (1), 3 × 10-9 mol/ml, and (c), thesensing interface that includes the double-stranded probe oligonu-cleotide–mutant (1) assembly after treatment with polymerase(Klenow fragment) and b-dCTP. Curve (d) shows the spectrum ofthe resulting surface after treatment with the avidin–alkaline phos-phatase conjugate, and curves (e) and (f) show the spectra of the bio-catalyzed precipitation of the insoluble product (5) for 10 min and40 min, respectively. The electron transfer resistance increases from1.6 kΩ to ∼ 4.1 kΩ upon the formation of the double-strandedassembly between the probe (3) and the mutant (1) oligonucleotides.This is consistent with the electrostatic repulsion of the redox labelfrom the electrode interface by the formation of the double-strandedassembly, thereby introducing a barrier for interfacial electron trans-fer. The surface coverage21 of the sensing interface by the target DNA(1), corresponds to 58%. Although the treatment of the electrodewith the polymerase and b-dCTP does not affect the interfacial elec-tron transfer resistance, the association of the hydrophobicavidin–alkaline phosphatase conjugate introduces a barrier for elec-tron transfer, Ret ≈ 6.2 kΩ. Biocatalyzed precipitation of (5) onto theelectrode insulates the electrode, a process that increases the electrontransfer resistance at the electrode. The electron transfer resistanceincreases to 8.4 kΩ and 16 kΩ upon the precipitation of (5) for 10and 40 min, respectively.

The Faradaic impedance spectra corresponding to the similarexperiments performed with the oligonucleotide representing thenormal sequence (2) are shown in Figure 2B. It is evident that after theformation of the double-stranded assembly, Ret = 3.9 kΩ, no increasein the electron transfer resistance is observed upon treatment of thesurface with avidin–alkaline phosphatase or an attempt to precipitatethe insoluble derivative (5). Thus, the successful analysis of (1) isattributed to the specific polymerase-mediated coupling of b-dCTP tothe mutant assembly, resulting in the biocatalytic precipitation of (5).

Figure 2C shows the Faradaic impedance spectra resulting uponthe analysis of a lower concentration of the mutant (1) correspondingto 1 × 10-12 mol/ml, following the steps outlined in Figure 1. The elec-tron transfer resistance is increased only from Ret = 1.6 kΩ to 2.0 kΩupon the formation of the double-stranded assembly. We attributethe slight increase in the electron transfer resistance at the electrode tothe low coverage of the surface with the oligonucleotide–(1) complex.This low coverage with the double-stranded assembly is, however,amplified upon the biocatalyzed precipitation of (5) by the

RESEARCH ARTICLE

nature biotechnology • VOLUME 19 • MARCH 2001 • http://biotech.nature.com254

Figure 1. Scheme for the electronic sensing of a single-base mutation usingthe biocatalytic precipitation of an insoluble product on the transducer.

Figure 2. Amplified detection of the DNA mutant (2) by the biocatalyzed precipitation of (5)using Faradaic impedance spectroscopy. (A) Faradaic impedance spectra (Zim vs. Zre)corresponding to (a) the (3)-modified electrode itself and (b) upon interaction with (1) (3 × 10-9 mol/ml) for 90 min. Curves (c)–(f): the double-stranded functionalized electrodes (c)after interaction with the Klenow fragment for 60 min; (d) upon the interaction of the biotin-labeled double-stranded assembly with avidin–alkaline phosphatase for 15 min; (e) after theinteraction of the avidin–alkaline phosphatase-labeled assembly with (4) for 10 min; (f) after theinteraction of the avidin–alkaline phosphatase-labeled assembly with (4) for 40 min. (B)Faradaic impedance spectra (Zim vs. Zre) corresponding to (a) the (3)-functionalized electrodeitself and (b) upon interaction with (2) (3 × 10-9 mol/ml). Curves (c)–(f): Repetition of the stepsoutlined in (A). (C) Faradaic impedance spectra (Zim vs. Zre) corresponding to (a) the (3)-functionalized electrode itself and (b) after interaction with (1) (1 × 10-12 mol/ml). Curves (c)–(f):Repetition of the steps detailed in (A). Conditions for the hybridization, polymerization, assemblyof the proteins, and the biocatalytic precipitation are detailed in the Experimental Protocol.

A B C

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://b

iote

ch.n

atu

re.c

om

© 2001 Nature Publishing Group http://biotech.nature.com

RESEARCH ARTICLE

avidin–alkaline phosphatase conjugate. The electron transfer resis-tance increases to 3.8 kΩ and 5.0 kΩ, upon the precipitation of theinsoluble product on the electrode for 10 and 40 min, respectively. Itshould be noted that the surface coverage of the probe oligonu-cleotide (3) on the electrode corresponding to 2.3 × 10-11 mol/cm2

represents the optimal surface density for efficient hybridization.Figure 3 shows the calibration curve that corresponds to changes

in the electron transfer resistances caused by the precipitation of(5) as a result of interactions of the sensing interface with differentconcentrations of the target analyte (1) (∆Ret corresponds to the dif-ference between the electron transfer resistance of the electrode afterthe precipitation of (5) and the electron transfer resistance of theelectrode after the hybridization with (1)). For the lowest sensitivitylimit, corresponding to 1 × 10-11 M (1 × 10-14 mol/ml), the signal-to-noise ratio is >3. (The noise level is assumed to be the system'sresponse upon analyzing (2), at 3 × 10-9 mol/ml, followed by anattempt to induce the precipitation of (5) for 40 min.) The resultspresented in Figure 3 show excellent reproducibility over the entireconcentration range. For a series of six electrodes in different experi-ments, at the lowest sensitivity, the different ∆Ret values were in therange of ∆Ret = 1.2 ± 0.1 kΩ. It should be noted that the detection ofthe single-base mismatch in (1) reveals high specificity and is notinfluenced by DNA contaminants. Analyzing (1) at the lower sensi-tivity limit (1 × 10-14 mol/ml) in the presence of denatured or dou-ble-stranded calf thymus DNA (1 × 10-9 M) yields ∆Ret values corre-sponding to 1.2 ± 0.2 kΩ, identical to the value observed for (1) inthe absence of contaminants. Also, analysis of the contaminantsusing this approach yield ∆Ret values corresponding to only 0.1–0.15 kΩ, implying that the contaminants do not affect the analysis ofthe mutant.

The specific detection of the single-point mutation in (1) by thebiocatalyzed precipitation of an insoluble product is further support-ed by the reaction of the complex formed between the probe oligonu-cleotide (3) and an oligonucleotide representing the normal sequence(2) with biotinylated uridine triphosphate (b-dUTP) in the presenceof polymerase, Klenow fragment (Fig. 4A). Covalent linkage of thebiotinylated base followed by the association of the avidin–alkalinephosphatase results in the formation of the insoluble product on theelectrode support, which is reflected in the high electron transferresistance, Ret = 19 kΩ (after 40 min of precipitation). Conversely, theinteraction of the complex formed between the probe oligonu-cleotide (3) and the mutant (1), with b-dUTP in the presence of poly-merase, followed by an attempt to precipitate the insoluble product(5), does not yield any increase in the electron transfer resistance (Fig. 4B), implying that no precipitation occurred on the electrode.

The analysis of the single-base mutant by the protocol outlined inFigure 1 is also demonstrated by microgravimetric quartz-crystal

microbalance assay of the biocatalyzed formation of the precipitate.The frequency of the piezoelectric crystal is controlled by the mass ofthe crystal. Thus, any increase in the mass associated with the crystal(∆m), for example as a result of the biocatalyzed precipitation of (5),will be accompanied by a decrease in the resonance frequency of thecrystal, –∆f. Gold-quartz crystals (9 MHz, AT-cut) were modifiedwith the thiolated oligonucleotide probe (3). The resulting oligonu-cleotide-functionalized crystals were then treated with a high con-centration (9 × 10-9 mol/ml) of the mutant (1), low concentration (1 × 10-12 mol/ml) of the mutant (1), or a high concentration (3 × 10-9 mol/ml) of the normal sequence (2). The resulting crystalswere then reacted with b-dCTP and polymerase Klenow fragment tocouple the biotinylated base to the assemblies that included the single-base mutation. Figure 5 (curves a and b) shows the time-dependent frequency changes of the probe oligonucleotide/(1)-functionalized crystal after the reaction with b-dCTP/polymerase and upon the interaction with avidin–alkaline phos-phatase conjugate, whereas curve c shows the frequency changes ofthe probe oligonucleotide/(2) interface after treatment with b-dCTP/polymerase and interaction with avidin–alkaline phos-phatase. The frequency changes of the crystal upon the biocatalyzedprecipitation of the insoluble product (5) on to the resulting assem-blies are shown in curves (d) (e), and (f), respectively. Treatment ofthe probe oligonucleotide/(1) mutant-functionalized crystal withavidin–alkaline phosphatase results in a frequency change of∆f = –47 Hz, curve (a), indicating the association of the conjugate tothe coupled b-dCTP base. The surface-associated enzyme conjugatebiocatalyzes the precipitation of (5), and an additional time-depen-dent decrease in the crystal frequency (∆f = about –70 Hz) isobserved, curve (d). Treatment of the probe oligonucleotide/(2)-functionalized crystal with avidin–alkaline phosphatase does notyield any frequency change, and no precipitation is observed uponaddition of (4), curves (c) and (f), respectively. The fact that b-dCTPis not coupled to the probe oligonucleotide/(2) assembly eliminates

http://biotech.nature.com • MARCH 2001 • VOLUME 19 • nature biotechnology 255

Figure 3. Calibration curve corresponding to changes in the electrontransfer resistances of the electrode as a result of the precipitation of (5) upon detection of different concentrations of (1). ∆Ret corresponds tothe difference in the electron transfer resistance as a result ofprecipitation of (5) and the electron transfer resistance after hybridizationof (1) with the sensing interface. Hybridization and precipitation wereperformed as described in the Experimental Protocol.

Figure 4. (A) Faradaic impedance spectra (Zim vs. Zre) corresponding to(a) the (3)-functionalized electrode itself and (b) after treatment with 3 × 10-9 mol/ml of normal oligonucleotide (2). Curves (c)–(f): (c) thedouble-stranded assembly after reacting with biotinylated dUTP andpolymerase; (d) the biotin-labeled double-stranded assembly afterinteraction with the avidin–alkaline phosphatase conjugate; (e) theassembly after biocatalyzed precipitation of (5) for 10 min; (f) theassembly after biocatalyzed precipitation of (5) for 40 min. (B) Faradaicimpedance spectra (Zim vs. Zre) corresponding to (a) the (3)-functionalizedelectrode itself and (b) after reacting with 3 × 10-9 mol/ml of the mutantoligonucleotide (1). Curves (c)–(f): Repetition of the steps outlined in (A).Conditions for the hybridization, assembly of the proteins, and thebiocatalytic precipitation are detailed in the Experimental Protocol.

A

B

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://b

iote

ch.n

atu

re.c

om

© 2001 Nature Publishing Group http://biotech.nature.com

the association of the enzyme conjugate and prohibits the subse-quent precipitation of (5). Treatment of the detector interface withthe lower concentration of the mutant results in smaller frequencychanges upon interaction with the avidin–alkaline phosphatase con-jugate and the subsequent precipitation of the insoluble product,curves (b) and (e), respectively. These results can be explained by theobservation that a low bulk concentration of (1) correlates with alow surface coverage of the mutant on the detector interface.

The use of polymerase-induced coupling of a biotinylated base,complementary to the mutation site, and the subsequent precipita-tion of the insoluble product by the linked avidin–alkaline phosphatase conjugate, is a general method to identify single-point-mutation sites. For example, the adenine base in position 15 ofthe 27-mer oligonucleotide (6) (5′-AGCGTAGCATAGATATACG-GTTCGCGC-3′) could be detected by the application of theoligonucleotide (7) (5′-HS-(CH2)6-GCGCGAACCGTATA-3′) as theprobe on the electrode. The double-stranded complex between theprobe oligonucleotide (7) and (6) that was formed on the electrodewas reacted with biotinylated uridine triphosphate, b-dUTP/poly-merase, to couple the biotin-labeled base to the assembly.Subsequent linking of avidin–alkaline phosphatase and biocatalyzedprecipitation of (5) onto the electrode results in an interfacial elec-tron transfer resistance of Ret = 23.0 kΩ. The interfacial electrontransfer resistance of the electrode is not altered when the double-stranded assembly between (6) and (7) is reacted with b-dCTP/polymerase, indicating that the specific coupling of the biotinylateduridine base results in the secondary biocatalyzed precipitation ofthe insoluble product on the transducer. It should be noted that forall positive mutant sequences, an increase in the electron transferresistance corresponding to ∆Ret ≥ 1.1 kΩ was observed at the lowestdetection limit, 1 × 10-11 M, whereas in the negative control experi-ments the interfacial electron transfer resistance increased only by∆Ret ≤ 0.2 kΩ at high DNA concentrations (3 × 10-6 M).

One major challenge to the method pertains to the adaptivity ofthe procedure to detect selectively the respective mutations in realclinical genomic samples22,23. Accordingly, we have analyzed threeblood samples that include the heterozygotic gene (sample I) and thehomozygotic gene (sample II), corresponding to the Tay–Sachsgenetic disorder, and the normal gene (sample III). The Tay–Sachsdisease is caused by a deficiency of the enzyme hexosaminidase,which degrades GM2 ganglioside to GM3 (ref. 24). The disease

appears at about six months of age and is fatal, usually in early child-hood. Affected children become blind, and physically and mentallyregressed. The disease is frequent in Ashkenazi Jews of easternEuropean descent. A recent survey25 reported that 1 in 30 AshkenaziJews is a carrier of the defective gene. A four-base insertion in theexon 11 (8), that encodes the α-chain of β-hexosaminidase is themost frequent mutant (70% of the carriers).

The sequences of the Tay–Sachs mutant and normal gene aredepicted in (8) and (8a), respectively (Table 1). The probe (9), which iscomplementary to the genes as far as one base preceding the mutation,was immobilized on gold electrodes. Table 1 shows the changes in theinterfacial electron transfer resistances of the (9)-functionalized goldelectrode as a result of the separate hybridization of different elec-trodes with samples I–III followed by the coupling of biotinylateddUTP and the biocatalyzed precipitation of (5). Clearly, only the het-erozygotic (sample I) and homozygotic genes (sample II) lead to highinterfacial electron transfer resistances, whereas the sample of the nor-mal gene yields a minute change in the interfacial electron transferresistance (∆Ret < 0.1 kΩ). To distinguish between the heterozygoticand homozygotic genes, the (9)-functionalized electrodes werehybridized separately with the three samples and reacted with thebiotinylated dCTP in the presence of polymerase. As expected, the het-erozygotic sample and the normal gene lead to the precipitation of theinsoluble product, whereas the homozygotic gene sample does notlead to any change in the interfacial electron transfer resistance (Table1). It should be noted that the analyzed genes were isolated from 0.5 mlof blood samples with no PCR amplification (see ExperimentalProtocol). Thus, the detector interfaces reveal impressive selectivity inanalyzing the target mutants within an enormous mixture of genomicfragments. Furthermore, the sensitivity of the method reveals the abil-ity to detect the mutant with no PCR amplification.

In conclusion, we have designed a bioelectronic method for thespecific detection of single-base mutations in DNA. The biocatalyticprecipitation of an insoluble product on the transducer provides anamplification route for the formation of the complex between thesensing interface and the mutant DNA. Faradaic impedance spec-troscopy and microgravimetric quartz-crystal microbalance mea-surements are effective electronic transduction methods for detectionof hybridization events on the surfaces. It should be noted, however,that the magnitudes of the transduced electronic signals are con-trolled by the surface coverage of the primary sensing interface andthe efficiency of hybridization (influenced by the specific basesequence and target length). The reported method is valid for the elu-cidation of a known single-base mismatch and the magnitude of thetransduced electronic signal should always be referenced to the nor-mal sequence response. The method, in its present configuration, canonly identify one of the four possible base polymorphisms on a singleelectrode. Future experiments that employ electrode arrays and pre-designed nucleic acid probes could enable the simultaneous detectionof the different genes of a specific genetic disorder, or eventually, the

RESEARCH ARTICLE

nature biotechnology • VOLUME 19 • MARCH 2001 • http://biotech.nature.com256

Figure 5. Time-dependent frequency changes upon the association ofavidin–alkaline phosphatase and the biocatalyzed precipitation of (5) inthe presence of the oligonucleotide/DNA assemblies: (a) and (b)correspond to the out-of-cell interaction of the (3)-functionalized electrodewith (1), 3 × 10-9 mol/ml and 1 × 10-12 mol/ml, respectively, followed by thereaction of the double-stranded functionalized assemblies with Klenowfragment and biotinylated dCTP, and in-cell recording of the frequencychanges of the resulting functionalized electrodes upon interaction withavidin–alkaline phosphatase, 100 nmol/ml. (c) Out-of-cell interaction ofthe (3)-functionalized gold-quartz crystal with (2), 3 × 10-9 mol/ml,followed by the reaction of the interface with the Klenow fragment andbiotinylated dCTP, and in-cell monitoring of frequency changes upon theinteraction of the interface with avidin–alkaline phosphatase, 100 nmol/ml. Curves (d) (e), and (f) correspond to the time-dependentfrequency changes of the respective interfaces formed in (a) (b), and (c)in the presence of (4), upon the biocatalyzed precipitation of (5).

Table 1. Change in the interfacial electron transfer resistance(∆Ret) upon the detection of (8) and (8a) by the biocatalyzed pre-cipitation of (5)a

Sample Reaction with Reaction with biotinylated biotinylated dUTP (∆Ret, kΩ) dCTP (∆Ret, kΩ)

(I) Heterozygote (8) 1.8 ± 0.1 1.6 ± 0.1(II) Homozygote (8) 2.3 ± 0.1 < 0.1(III) Normal (8a) < 0.1 2.1 ± 0.1

aSequences are as follows:(8) ....3′-GAGACGGGGGACCATGGACTTGGCATATAGATAGGATTAC-5′....(8a) ....3′-GAGACGGGGGACCATGGACTTGGCATATAGGATACCGGG-5′....(9) 5′-HS(CH2)6CTCTGCCCCCTGGTACCTGAACCGTATATC-3′

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://b

iote

ch.n

atu

re.c

om

© 2001 Nature Publishing Group http://biotech.nature.com

RESEARCH ARTICLE

mutants of several genetic disorders. In contrast to the available tech-niques for detecting single-point mutations by comparative assays ofthe signals originating from DNA with two or three-base mismatchesat identical concentrations, our method yields a defined positive ornegative transduction signal over a broad sensitivity range. Thus thesingle-base mismatch may be detected in a DNA mixture indepen-dently of its relative concentration. Also, the method described here issensitive, and enables the quantitative analysis of nucleic acids withno PCR pre-amplification.

Experimental protocol Electrochemical measurements were performed with an electrochemicalimpedance analyzer (Model 1025; EG&G, Oak Ridge, TN) and a potentiostat(Model 283; EG&G) connected to a computer (Software Power Suite 1.03;EG&G). Electrochemical measurements were performed in a three-electrodecell consisting of the DNA-modified electrode, a glass carbon auxiliary elec-trode isolated by a glass frit, and saturated calomel electrode (SCE). All imped-ance measurements were performed in 0.1 M phosphate buffer (pH 7.4), whichincluded 5 × 10-3 M K3[Fe(CN)6]: K4[Fe(CN)6], in a 1:1 ratio, as a redox probe.An alternating voltage (10 mV) was applied at a bias potential of 0.180 V in thefrequency range 100 MHz–20 kHz. Microgravimetric quartz-crystal microbal-ance (QCM) experiments were done with a frequency analyzer (Fluke model163/164; Almelo, The Netherlands) linked to a personal computer. Quartz crys-tals (9 MHz, AT-cut) sandwiched between two gold electrodes (area 0.196 cm2,roughness factor ∼ 3.5) were used in the microgravimetric experiments.

The thiolated oligonucleotides were freshly prepared before the modifica-tion of the electrodes, by dithiothreitol (DTT) reduction of the commerciallyordered respective disulfides. The disulfide was dissolved in 0.5 ml of 0.17 Mphosphate buffer pH 8.0, and treated with 0.04 M DTT for 16 h at room tem-perature. The thiolated DNA was purified by elution of the DNA through aNAP-10 column of Sephadex G-25 (Amersham Pharmacia Biotech,Buckinghamshire, England) with 0.17 M PBS, to yield a DNA solution of∼ 0.5 ml (4.8 µM). Immobilization of the probe thiolated oligonucleotides onthe gold surfaces was done by incubation of the clean gold electrodes with thethiolated oligonucleotide, 4.8 µM in PBS for 12 h. Hybridization of the sens-ing interface with the analyte DNA (mutant or normal gene) was done in 2× SSC (sodium saline citrate) buffer (pH 7.5) for 90 min at 25°C.

Coupling of the biotinylated label to the double-stranded DNA was doneby treatment of the functionalized electrode with biotinylated dCTP, ordUTP (20 nmol/ml) in the presence of Klenow fragment of polymerase I

(Sigma, Rehovot, Israel), 20 U/ml in Tris buffer solution 20 mM (pH 7.8),which included 10 mM MgCl2, 60 mM KCl, and 1 mg/ml Tris-(2-car-boxyethyl)phosphine at 37°C for 60 min.

The avidin–alkaline phosphatase conjugate was coupled to the biotinylat-ed surface by treating the electrodes with avidin–alkaline phosphatase (100 nmol/ml) in 0.1 M phosphate buffer for 15 min. The electrodes with thelinked avidin–alkaline phosphatase were interacted with 5-bromo-4-chloro-3-indoyl phosphate (4), in 0.1 M Tris buffer solution (pH 7.6) for variabletimes to induce the biocatalyzed precipitation of (5) on the transducers.

For the analysis of the Tay–Sachs genetic disorder in a genomic sample, therespective DNAs were separated from blood samples. The blood samples (0.5 ml) were washed with a lysis buffer (PBS, pH 7.4, 0.2% Nonidet P-40)and centrifuged at 500 g at room temperature for 8 min. The resulting pelletwas incubated for 24 h at 37°C in 1 ml of digestion buffer (10 mM Tris-HCl,pH 8.0, 100 mM NaCl, 25 mM ethylenediamine tetraacetate (EDTA), 0.5%sodium dodecyl sulfate, 0.1 mg proteinase K). The DNA was subsequentlyextracted by the phenol method and solubilized in Tris-HCl/EDTA buffer.The concentration of DNA in sample I was 74 ng/µl, in sample II 86 ng/µl,and in sample III 154 ng/µl. The DNA samples were digested for 30 min atroom temperature with DNase (Sigma), 1.6 mU in 20 µl of 0.1 M sodiumacetate, 5 mM MgSO4, pH 5.0. The process was terminated by the addition ofEDTA, 20 mM. The resulting DNA was purified by the concert rapid plasmidpurification kit (Gibco BRL, Gaithersburg, MD). The final concentrations inthe purified samples I, II, and III were 15, 18, and 31 µg/ml, respectively. Thesamples were denatured by incubation at 90°C for 10 min followed by rapidquenching in an ice bath. The volumes of the samples were adjusted with 6× SSPE (saline–sodium phosphate–EDTA) buffer (pH 7.4) to include 8 ng/µl each for samples I and II, and 16 ng/µl for sample III. The (9)-functionalized electrodes were hybridized with the respective DNA samplesby incubating the electrodes in the samples for 60 min at 5°C, followed byincubation at 45°C for 90 min. The resulting electrodes were washed with aTris-HCl buffer solution, pH 7.4. The resulting electrodes were reacted withthe biotinylated dUTP or biotinylated dCTP nucleotide, followed by the bio-catalyzed precipitation of (5), as described above.

AcknowledgmentsParts of the research are supported by the Max Planck Research Award forInternational Cooperation and by the Israel Ministry of Science as anInfrastructure Project in Biomicroelectronics.

Received 13 April 2000; accepted 20 December 2000

http://biotech.nature.com • MARCH 2001 • VOLUME 19 • nature biotechnology 257

1. Winzler, E.A. et al., Direct allelic variation scanning of the yeast genome. Science281, 1194–1197 (1998).

2. Boon, E.M., Ceres, D.M., Drummond, T.G., Hill, M.G. & Barton, J.K., Mutation detec-tion by electrocatalysis at DNA-modified electrodes. Nat. Biotechnol. 18, 1096–1100(2000).

3. Mikkelsen, S.R., Electrochemical biosensors for DNA sequence detection.Electroanalysis 8, 15–19 (1996).

4. Rodriguez, M. & Bard, A.J., Electrochemical studies of the interaction of metalchelates with DNA. 4.Voltammetric and electrogenerated chemiluminescent studiesof the interaction of tris(2,2'-bipyridine)osmium(II) with DNA. Anal. Chem. 62,2658–2662 (1990).

5. Adleman, L.M. Molecular computation of solutions to combinatorial problems.Science 266, 1021–1024 (1994).

6. Liu, Q.H., Frutos, A.G., Thiel, A.J., Corn, R.M. & Smith, L.M., DNA computing on sur-faces: encoding information at the single base level. J. Comput. Biol. 5, 269–278(1998).

7. Frutos, A.G., Smith, L.M. & Corn, R.M., Enzymatic ligation reactions of DNA 'words'on surfaces for DNA computing. J. Am. Chem. Soc. 120, 10277–10282 (1998).

8. Jordan, C.E., Frutos, A.G., Thiel, A.J.& Corn, R.M., Surface plasmon resonance imag-ing measurements of DNA hybridization adsorption and streptavidin/DNA multilayerformation at chemically modified gold surfaces. Anal. Chem. 69, 4939–4947 (1997).

9. Korri-Youssoufi, H., Granier, F., Srivastava, P., Godillot, P. & Yassar, A., Towards bio-electronics: specific DNA recognition based on an oligonucleotide- functionalizedpolypyrrole. J. Am. Chem. Soc. 119, 7388–7389 (1997).

10. Millan, K.M., Saraullo, A. & Mikkelsen, S.R., Voltammetric DNA biosensor for cystic-fibrosis based on a modified carbon-paste electrode. Anal. Chem. 66, 2943–2948(1994).

11. Hashimoto, K., Ito, K. & Ishimori, Y., Sequence-specific gene detection with a goldelectrode modified with DNA probes and an electrochemically active dye. Anal.Chem. 66, 3830–3833 (1994).

12. Bardea, A., Patolsky, F., Dagan, A. & Willner, I., Sensing and amplification of oligonu-cleotide–DNA interactions by means of impedance spectroscopy: a route to aTay–Sachs sensor. Chem. Commun. pp. 21–22 (1999).

13. Wang, J. et al., Peptide nucleic acid probes for sequence-specific DNA biosensors. J.

Am. Chem. Soc. 118, 7667–7670 (1996).14. Bardea, A., Dagan, A., Ben-Dov, I., Amit, B. & Willner, I., Amplified microgravimetric

quartz crystal microbalance analysis of oligonucleotide complexes: a route toTay–Sachs biosensor device. Chem. Commun. pp. 839–840 (1998).

15. Caruana, D.J. & Heller, A. Enzyme amplified amperometric detection of hybridizationand of a single base pair mutation in an 18-base oligonucleotide on a 7-µm diametermicroelectrode. J. Am. Chem. Soc. 121, 769–774 (1999).

16. Land, H., Parada, L.F. & Weinberg, R.A., Cellular oncogenes and multistep carcino-genesis. Science 222, 771–777 (1983).

17. Almoguera, C. et al., Most human carcinomas of the exocrine pancreas containmutant c-K-ras. Cell 53, 549–554 (1988).

18. Watanabe, H. et al. K-ras mutations in duodenal aspirate without seretin stimulationfor screening of pancreatic and biliary tract carcinoma. Am. Cancer Soc. 86,1441–1448 (1999).

19. Patolsky, F., Zayats, M., Katz, E. & Willner, I., Precipitation of an insoluble product onenzyme monolayer electrodes for biosensor applications: characterization byFaradaic impedance spectroscopy, cyclic voltammetry and microgravimetric quartzcrystal microbalance analyses. Anal. Chem. 71, 3171–3180 (1999).

20. Patolsky, F., Katz, E., Bardea, A. & Willner, I., Enzyme linked amplified electrochemi-cal sensing of oligonucleotide–DNA interactions by means of precipitation of aninsoluble product and using impedance spectroscopy. Langmuir 15, 3703–3706(1999).

21. Steel, A.B., Herne, T.M. & Tarlov, M.J., Electrochemical quantification of DNA immo-bilized on gold. Anal. Chem. 70, 4670–4677 (1998).

22. Schafer, A.J. & Hawkins, J.R. DNA variation and the future of human genetics. Nat.Biotechnol. 16, 33–39 (1998).

23. Cargill, M. et al. Characterization of single-nucleotide polymorphisms in codingregions of human genes. Nat. Genet. 22, 231–238 (1999).

24. Gravel, R.A. et al. The GM2 gangliosides. In The metabolic and molecular bases ofinherited diseases, Vol. 2. (eds Scriver, C.R., Beaudet, A.L., Sly, W.S. & Valle, D.)2839–2879 (McGraw-Hill, New York; 1995).

25. Myerowitz, R. & Costigan, F.C., The major defect in Ashkenazi Jews with Tay–Sachsdisease is an inversion in the gene for the α-chain of β-hexosaminidase. J. Biol.Chem. 263, 18587–18589 (1988).

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://b

iote

ch.n

atu

re.c

om

© 2001 Nature Publishing Group http://biotech.nature.com