Biosensors and Bioelectronics 19 (2003) 165�/175

www.elsevier.com/locate/bios

Rapid electrochemical genosensor assay using a streptavidin carbon-polymer biocomposite electrode

E. Williams 1, M.I. Pividori 2, A. Merkoci, R.J. Forster 1, S. Alegret *

Grup de Sensors i Biosensors, Departament de Quımica, Universitat Autonoma de Barcelona, 08193 Bellaterra, Catalonia, Spain

Received 9 May 2002; received in revised form 10 April 2003; accepted 23 April 2003

Abstract

A sensor capable of detecting a specific DNA sequence was designed by bulk modification of a graphite epoxy composite

electrode with streptavidin (2% w/w). Streptavidin is used to immobilise a biotinylated capture DNA probe to the surface of the

electrode. Simultaneous hybridisation occurs between the biotin DNA capture probe and the target-DNA and between the target-

DNA and a digoxigenin modified probe. The rapid binding kinetic of streptavidin�/biotin allows a one step immobilisation/

hybridisation procedure. Secondly, enzyme labelling of the DNA duplex occurs via an antigen�/antibody reaction between the Dig-

dsDNA and an anti-Dig-HRP. Finally, electrochemical detection is achieved through a suitable substrate (H2O2) for the enzyme-

labelled duplex. Optimisation of the sensor design, the modifier content and the immobilisation and hybridisation times was attained

using a simple nucleotide sequence. Regeneration of the surface is achieved with a simple polishing procedure that shows good

reproducibility. The generic use of a modified streptavidin carbon-polymer biocomposite electrode capable of surface regeneration

and a one step hybridisation/immobilisation procedure are the main advantages of this approach. In DNA analysis, this procedure,

if combined with the polymerase chain reaction, would represent certain advantages with respect to classical techniques, which prove

to be time consuming in situations where a simple and rapid detection is required. This innovative developed material may be used

for the detection of any analyte that can be coupled to the biotin�/streptavidin reaction, as is the case of immunoassays.

# 2003 Elsevier B.V. All rights reserved.

Keywords: Streptavidin�/biotin; Streptavidin carbon-polymer biocomposite; Horseradish peroxidase labelling; Amperometric genosensor; DNA

hybridisation; Digoxigenin; Anti-digoxigenin antibody

1. Introduction

‘World-wide infectious diseases account for approxi-

mately 40% of the total 50 million annual deaths, with

microbial diseases accounting for the majority of deaths in

developing countries ’ (World Health Organisation offi-

cial statistics 2000). Clearly, the development of a rapid,

sensitive, selective and highly accurate method for the

detection of DNA is one of the key health issues facing

our society today. Two competitive approaches exist for

DNA detection: direct sequencing and DNA hybridisa-

* Corresponding author. Tel.: �/34-93-581-1976; fax: �/34-93-581-

5972.

E-mail address: salegret@gsb.uab.es (S. Alegret).1 Dublin City University, Glasnevin, Dublin 11, Ireland.2 Facultad de Bioquımica y Ciencias Biologicas, Universidad

Nacional del Litoral, Santa Fe, Argentina.

0956-5663/03/$ - see front matter # 2003 Elsevier B.V. All rights reserved.

doi:10.1016/S0956-5663(03)00171-4

tion. Although DNA sequencing has proved indispen-

sable for resolving the genetic make up of many

pathogenic species, as an analytical technique is too

time consuming (Turner et al., 1997). DNA hybridisa-

tion proves to be a simpler, quicker and more selective

method for the detection of previously known DNA

sequences.

The possibility of direct electrochemical detection of

DNA was initially proposed by Palecek (1958, 1960),

who recognised the capability of both DNA and RNA

to yield reduction and oxidation signals.Advancements in the areas of genetics with regards to

the elucidation of specific genes responsible for various

pathogenic diseases have encouraged research in the

area of genosensor development. This advancement is

shown by a large surge of publications relating to

electrochemical DNA detection as recently reviewed by

Pividori et al. (2000) and Palecek and Fojta (2001).

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175166

While many procedures still rely on the oxidation of the

guanine residues (Armistead and Thorp, 2001; Burrows

and Muller, 1998), much work has been done in

perfecting the immobilisation of DNA onto an electrodesurface along with the incorporation of electroactive

markers.

Extensive studies by Barton and co-workers (Erkkila

et al., 1999; Rajksi et al., 2000) have demonstrated the

capability of certain compounds to interact either by

intercalation, hydrophobic binding to the major and

minor grooves of DNA, or electrostatic interaction with

dsDNA, such as Ru(II) (Armistead and Thorp, 2000)Khellin (Radi, 1999) or Co(phen)3

3� (Li et al., 1997).

While, such a phenomena is of great interest in terms of

drug delivery, as in the case of daunomycin, it does not

lend itself easily for the development of genosensors due

to problems of non-specific adsorption onto the elec-

trode surface. Additionally, it has been reported (Erk-

kila et al., 1999) that the extent of intercalation depends

on the DNA sequence, favouring guanine rich DNAsequences. Such favouritism limits the range of its

application.

A more frequent use of electroactive markers involves

their attachment to the DNA duplex by direct covalent

attachment to ssDNA, utilisation of the biotin�/strepta-

vidin chemistry (Alfonta et al., 2001; Pividori et al.,

2001a,b) and the digoxigenin�/antidigoxigenin reaction

(Lumley-Woodyear et al., 1999). Ferrocene derivativesare commonly used as electroactive indicators by

immobilising them to a DNA probe (Xu et al., 2001;

Takenaka et al., 1994; Uto et al., 1997). More recent

studies by Kizek et al. (2002) have shown how using

osmium labelled ssDNA results in detection limits in the

region of 5�/10 ng/ml. Other studies show the labelling

of DNA with enzymes such as peroxidase, a known

redox enzyme which due to its slow heterogeneouselectron transfer kinetics require an additional electron

mediator to ascertain the response. Enzymes that have

been incorporated are horseradish peroxidase, (Pividori

et al., 2001a,b; Alfonta et al., 2001; Chen et al., 2000;

Lumley-Woodyear et al., 1999) and alkaline phospha-

tase (Palecek et al., 2002).

Various methods for the attachment of ssDNA

probes to the surface of the electrode exist. In brief,these methods may involve the use of an electrochemi-

cally pre-treated carbon electrode resulting in an in-

crease in the hydrophilicity of the transducer.

Subsequently, holding the electrode at positive poten-

tials results in the adsorption of the DNA probe through

hydrophilic and electrostatic attraction (Wang et al.,

1996). Covalent attachment of ssDNA to the surface of

the electrode can result in strong stable linkage with ahigh surface coverage, a key feature for the development

of sensors. Numerous methods for the covalent attach-

ment of biocomponents to the surface of electrodes are

reported in literature (Lee and Shim, 2001; Millan et al.,

1992; Millan and Mikkelsen, 1993), resulting in an

activated electrode surface which is capable of binding

with ssDNA. Covalent bonding of ssDNA to gold

electrodes has been achieved by attaching a mercapto-

hexyl group to the 5?-phosphate end of a DNA strand

(Hashimoto et al., 1994, 1998; Herne and Tarlov, 1997).

The aim of these efforts is to produce self assembled

monolayers similar to those described elsewhere (Nuzzo

and Allara, 1983).

In most of the aforementioned techniques, (Wang et

al., 1996, 1997a,b; Wang and Kawde, 2001; Marrazza et

al., 1999; Chiti et al., 2001) immobilisation of the DNA

probe depends on the chemical or electrochemical pre-

treatment step of the transducer. Many systems have

been developed involving the immobilisation of biotin

on the surface of electrodes in order to use biotin�/

avidin�/biotin sandwich techniques. One such immobi-

lisation procedures was proposed by Pantano et al.

(1991). However, a chemical pre-treatment step is again

required to immobilise the biotin onto the electrode

surface. Further advances and manipulation of the

biotin�/avidin�/biotin chemistry have been achieved

using UV light to initiate the attachment of a photo-

active biotin molecule to a transducer surface (Sha-

mansky et al., 2001).

An attractive feature when developing sensors is the

possibility of surface regeneration for repeated use. In

order to achieve these objectives in the aforementioned

references it is necessary to denature the DNA duplex.

Wang et al. (1998) have overcome the necessity of a

denaturation step by using a ssDNA probe dispersed

uniformly in the matrix of a polishable and renewable

carbon paste electrode. Here, the carbon paste acted as a

reservoir for ssDNA probes, such that each nucleic acid

assay was performed on a new surface using the same

electrode. Again however, the surface of the electrode

had to be activated.

This report describes the bulk modification of a

carbon-polymer composite electrode with streptavidin.

Exploitation of the strong interaction between biotin�/

streptavidin allows immobilisation of a biotin labelled

capture probe, containing the DNA sequence comple-

mentary to the target probe. Hybridisation of the target-

DNA to the capture probe and an antigen labelled DNA

sequence is achieved simultaneously. Electrochemical

detection is achieved when an enzyme-labelled antibody,

specific to the antigen (digoxigenin) labelled DNA is

added. One of the main advantages of this proposed

genosensor design is the capability of surface regenera-

tion through a simple polishing procedure along with a

one step immobilisation and hybridisation procedure.

To the best of the author’s knowledge the use of

immobilised streptavidin within a carbon composite in

addition to a one step immobilisation/hybridisation

procedure has not been reported yet.

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175 167

2. Materials and instrumentation

2.1. Chemicals and solutions

The composite electrodes were prepared using gra-

phite powder with a particle size of 50 mm (BDH, UK),

Epotek H77 (epoxy resin) and a hardener (Epoxy

Technology, USA). The DNA oligomers were obtained

from TIB MOLBIOL (Germany). The base sequences of

these oligomers can be seen below:

. Capture probe, biotin�/DNA(A): biotin-5?-AAA

AAA AAA AAA AAA AAA AA

. Target-DNA(B): 5?-AAA AAA AAA AAA AAA

AAA AAA AAA ATT TTT TTT TTT TTT TTT

TTT TTT TT-3?. Digoxigenin�/DNA(C): Dig-3?-TTT TTT TTT TTT

TTT TTT TT-5?

The hybridisation solution (10�/ SSC, 2�/ Den-

hardt’s, 200 mg/ml chloroform extracted salmon testes

DNA), bovine serum albumin, sodium dodecyl sulphate,

Tween 20, hydroquinone, and HRP�/biotin (peroxidase

biotinamidocaproyl) were purchased from Sigma. The

anti-digoxigenin horseradish peroxidase (anti-Dig-

HRP) and streptavidin were purchased from Roche

(Germany). Tris�/HCl, NaCl and hydrogen peroxide

were purchased from Merck (Germany). All other

reagents were of the highest grade available. All aqueous

solutions were prepared using double distilled water.

Fig. 1. Schematic representation of the DNA analysis based on an

electrochemical streptavidin carbon-polymer biocomposite. (A) Strep-

tavidin modified composite electrode. (B) One step immobilisation/

hybridisation procedure: target-DNA is hybridised with the biotiny-

lated capture probe and with the digoxigenin modified probe. The

complex biotin�/dsDNA-Dig is immobilised on the electrode surface

by linking the biotinylated capture probe with the streptavidin present

in the biocomposite. (C) Enzyme labelling based on the immunological

reaction between the immobilised dsDNA-Dig with anti-Dig-HRP.

(D) Electrochemical detection of the enzyme (HRP) labelled dsDNA.

2.2. Instrumentation and electrochemical measurements

Amperometric measurements were done with a LC-

4C Amperometric Controller (BAS Bioanalytical Sys-

tem, USA). Cyclic voltammetry measurements were

carried out using Autolab PGSTAT 20 (Eco-chemie,

The Netherlands). A platinum auxiliary electrode (Cri-

son 52�/67 1) and a double junction Ag/AgCl reference

electrode (Orion 900200 with 0.1 M KCl as an external

reference solution) were used. The carbon-polymer

composite electrode modified with streptavidin made

in our laboratory was used as the working electrode.

The incubations at controlled temperature were carried

out in the Eppendorf Thermomixer 5436.

Both cyclic voltammetry and chronoamperometry

were done in a 20 ml electrochemical cell. The electrolyte

in all cases was 0.1 M phosphate buffer, pH 7.4,

containing 0.1 M KCl and 1.81 mM hydroquinone.

For cyclic voltammetry the potential window ranged

between �/0.75 and 1.25 V (vs. Ag/AgCl). For chron-

oamperometry the applied potential was �/0.250 V (vs.

Ag/AgCl).

2.3. Construction of the streptavidin carbon-polymer

biocomposite electrodes

Graphite powder and epoxy resin were mixed in a 1:4(w/w) ratio. For every gram of graphite/epoxy mixture,

an additional 20 mg of streptavidin were added*/

resulting in a 2% (w/w) streptavidin carbon-polymer

biocomposite. This mixture was thoroughly hand mixed

to ensure the uniform dispersion of the streptavidin and

the carbon throughout the polymer. The resulting paste

was placed in a PVC cylindrical body (6 mm i.d.) that

had an electrical contact placed at a depth of 3 mm. Thecomposite was cured at 40 8C for 1 week (Santandreu et

al., 1997). Prior to each use the biocomposite was wetted

with doubly distilled water and then polished. Firstly

with abrasive paper and then with alumina paper

(polishing strips 301044-001, Orion) to give a smooth

mirror finish with a fresh renewable surface.

3. Procedure

3.1. Immobilisation and hybridisation

Fig. 1 illustrates the experimental procedure consist-

ing of three steps. Firstly, to an eppendorf containing a

volume of 120 ml of the hybridisation solution, pre-

Fig. 2. Cyclic voltammograms of (A) the streptavidin carbon-polymer

biocomposite, and (B) the carbon-polymer composite electrode. Scan

rate was 0.05 V/s in 0.1 M phosphate buffer pH 7.4, with 0.1 M KCl

and 1.81 mmol hydroquinone. All potential are recorded vs. Ag/AgCl

reference electrode.

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175168

viously equilibrated to 42 8C, 10 ml of each of the three

oligonucleotides, mentioned in Section 2.1, were added.

The final amount of all three DNA strands in the

hybridisation solution is 150 pmole. To this eppendorf,a freshly polished streptavidin carbon-polymer biocom-

posite electrode (Fig. 1A) was immersed and shaken

gently for 60 min at 42 8C. Immobilisation of DNA

occurs through the biotinylated capture probe. Simulta-

neous hybridisation occurs between the biotinylated

capture probe and the target-DNA and between the

target-DNA and a Dig-DNA as seen in Fig. 1B.

Removal of the electrode from the eppendorf is followedby a gentle washing in the thermomixer at 42 8C for 5

min using 0.01 M Tris�/HCl at pH 7.4, thereby removing

all unbound DNA.

3.2. Enzyme labelling

The next stage involves the attachment of the enzyme

label, anti-Dig-HRP to Dig-dsDNA, Fig. 1C. This

attachment is achieved by adding 10 ml of anti-Dig-HRP (12.45 mg/ml) to an eppendorf containing a volume

of 140 ml of blocking solution (1�/ PBS, 2% w/v BSA,

0.1% Tween 20, 5 mM EDTA) (Kricka, 1992), pre-

viously equilibrated to 37 8C. To this eppendorf, the

DNA modified streptavidin carbon-polymer biocompo-

site electrode was immersed and shaken gently for 40

min at 37 8C. The electrode is then washed with a post-

incubation solution (10 mM sodium phosphate pH 6.5,0.5 M NaCl, 0.05% w/v Tween 20, 0.1% w/v BSA, 1 mM

EDTA) (Kricka, 1992), with slight agitation at 37 8C for

5 min using the thermomixer. Once the immobilisation,

hybridisation and enzyme labelling have been achieved

the electrodes if not required immediately were stored in

Tris buffer pH 7.4 at 4 8C thereby allowing electro-

chemical measurements can be carried out later on.

3.3. Electrochemical enzyme activity determination

The amount of enzyme-labelled antibody bound to

the Dig-DNA duplex was determined using the follow-

ing electrochemical detection procedure (Fig. 1D).

Experiments were carried out at 229/2 8C using the

hydroquinone electrolyte solution mentioned in Section

2.2. Hydroquinone buffer was prepared daily because ofits sensitivity to photo-oxidation, by adding the correct

quantity of hydroquinone to a 0.1 M phosphate buffer

at pH 7.4 and 0.1 M KCl. This buffer was purged with

nitrogen prior to use for at least 1 h. Chronoampero-

metric experiments were then performed at a potential

of �/0.250 V vs. Ag/AgCl electrode. This potential of �/

0.250 V was chosen from cyclic voltammetry experi-

ments Fig. 2A. For the enzymatic activity determina-tions the freshly prepared carbon biocomposite

containing the immobilised enzyme-labelled dsDNA

was immersed in the hydroquinone electrolyte solution

and allowed to equilibrate for 10 min. Subsequently,

H2O2 was added to a final concentration of 176 mM,

which corresponds to the H2O2 concentration capable to

saturate the whole enzyme amount employed in the

labelling procedure. At Fig. 1D and Eq. (1) are

represented the reactions that occur at the genosensor

surface polarised at �/0.250 V (vs. Ag/AgCl) uponH2O2 addition in the presence of hydroquinone. After

addition of the substrate, the current was again mea-

sured for a further 10 min. This time is chosen to ensure

that a steady state current had been reached maximising

the catalytic activity of the enzyme horseradish perox-

idase. The electrode response is defined as the difference

in current before and after the addition of H2O2.

HRP(red)�H2O2 0 HRP(ox)�H2O

HRP(ox)�HQ(red) 0 HRP(red)�HQ(ox)HQ(ox)�ne� 0 HQ(red)

(1)

3.4. Variation of streptavidin content in the carbon-

polymer biocomposite. Stability study

The optimisation of the streptavidin content in the

biocomposite was done using electrodes with fourdifferent streptavidin contents (1, 1.5, 2 and 5%). These

electrodes were evaluated for 3 weeks storing them in an

upright position at 4 8C.

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175 169

The stability of the streptavidin biocomposite electro-

des was also studied over a 12-week period, during

which time the electrodes were also stored in an upright

position at 4 8C.In all cases four electrodes were used, and the same

polishing procedure described in Section 2.3 was fol-

lowed.

For these evaluations, the procedure involved immer-

sing the streptavidin modified electrodes in a 150 ml

biotin�/HRP in excess concentration of 50 mg/ml for 60

min at 37 8C with gentle agitation again using the

thermomixer. It is known that streptavidin forms acomplex with biotin at room temperature with an

association constant in the range of 1015/M (Gonzalez

et al., 1999). Electrochemical detection was carried out

as described in Section 3.3 above for the detection of

HRP labelled DNA duplex.

3.5. Non-specific adsorption evaluation. Study of the

mechanism of the one step immobilisation/hybridisation

procedure

As with all methods that use an enzyme label, the

problem of non-specific adsorption needs to be ad-

dressed. Furthermore, knowledge of the one step

immobilisation and hybridisation mechanism is useful

to understand how the genosensor works.

For these studies, (i) carbon-polymer compositeelectrodes, which contains no streptavidin, were used

and the same procedure described above was followed.

In the other three assays (ii, iii and iv), the streptavidin

carbon-polymer biocomposite electrodes prepared as

described above were used. The same procedure was

followed but omitting (ii) the target-DNA and (iii) the

biotinylated capture probe. Finally, assay (iv) was done

using all three oligonucleotide components to evaluatethe genosensor in capture format. In all cases (assays i,

ii, iii and iv) four electrodes were used following the

same polishing procedure described in Section 2.3.

3.6. Optimisation of genosensor response

3.6.1. Immobilisation as the limiting step

This assay was done using a two step procedure.

Immobilisation times were varied from 10 to 120 min atregular intervals of 10 min. For subsequent hybridisa-

tion, a time of 60 min was chosen in agreement with

literature (Gingeras et al., 1987). Detection of the

hybridisation was achieved using the electrochemical

procedure described above.

3.6.2. Hybridisation as the limiting step

This assay was also done using a two step procedure.In this instance, a constant immobilisation time of 60

min was used. Again, a variation of the hybridisation

time from 10 to 120 min at regular intervals of 10 min

was chosen. All other experimental conditions were the

same as above.

3.6.3. Variation of time for the one step hybridisation/

immobilisation procedure

This final assay used a one step procedure. A study of

the experimental time for the one step hybridisation/

immobilisation procedure vs. current response was

done. Again, a variation of the one step proceduretime between 10 and 120 min at regular intervals of 10

min was chosen. All other experimental conditions are

the same as above.

In all experiments, four different electrodes were used

and the electrode surface was renewed before each

assay.

4. Results and discussion

4.1. Cyclic voltammetry of streptavidin modified

electrodes

Cyclic voltammograms of the 2% (w/w) streptavidin

modified composite electrode with immobilised DNA in

a 1.81 mM hydroquinone phosphate buffer solution are

shown in Fig. 2A. The purpose of cyclic voltammetry is

to establish the location of the reduction potential ofbenzoquinone (HQox). In Fig. 2A*/for streptavidin

carbon-polymer biocomposite electrode*/the upper

curve represents the oxidation of hydroquinone to

benzoquinone with an Ep at 0.537 V, while the lower

curve represents the reverse reaction with an Ep at �/

0.230 V. Fig. 2B shows the voltammetry for a carbon-

polymer composite (without streptavidin), an electrode

used in the non-specific adsorption studies. In thisinstance the location of Ep for the oxidation and

reduction peak potentials is at 0.249 and �/0.127 V,

respectively. The difference in the peak potentials

between both composite electrodes may be attributed

to the presence of streptavidin, which may affect the

conductivity, surface properties and therefore the loca-

tion of the formal potential (Cespedes et al., 1996). It is

known that in the presence of a partially blockedelectrode the value for the peak-to-peak separation

will increase, with the oxidation peak moving in a

more positive direction and the reduction peak moving

in a more negative direction. However, since in this

instance only 2% of the total surface area of the

electrode contains streptavidin available for binding

with a biotin-probe, any shift in potential would be

negligible if compared with a genosensor design byHerne and Tarlov (1997) that use a self assembled

monolayer with the intention of achieving complete

surface coverage.

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175170

4.2. Variation of streptavidin content in the carbon-

polymer biocomposite

From a previously published report (Alegret, 1996) it

is known that varying the biocomponent content in a

carbon-polymer biocomposite can adversely affect the

overall electrochemical performance of a given sensor.

This loss of electrochemical performance is attributable

to the increased concentration of modified species

resulting in a separation of the graphite particles,

causing a reduction of the conductivity. For this reason

it was decided to optimise the streptavidin content in the

composition. This was achieved by constructing bio-

composites with 1, 1.5 and 2% w/w of streptavidin. The

experiments were carried out as described previously,

over a 3-week period during which the electrodes were

stored in an upright position at 4 8C. In all experiments,

four different electrodes were used and polished to a

mirror finish prior to immobilising biotin�/HRP on the

transducer surface. From Fig. 3 it can be seen that all

three types of biocomposite electrodes decrease their

current response with time, this is an issue to be

addressed in the following section. Additionally, there

is a decrease in current response for electrodes having a

lower w/w of streptavidin available for reaction with the

biotin�/HRP. This data suggests that increasing the

percentage of streptavidin from 1 to 2% in the biocom-

posite does not adversely affect the electrochemical

response due to loss of conductivity. Additionally, an

electrode with 5% streptavidin was also prepared. Upon

immersion into the electrolyte solution this electrode

appeared to ‘dissolve’ possibly because the high bio-

component content prevents sufficient crosslinking of

the polymer to provide a rigid matrix.

Fig. 3. Current response of carbon-polymer biocomposite electrodes

with different streptavidin content (SA). Experiments were done in 0.1

M phosphate buffer, 0.1 M KCl, pH 7.4 with 1.81 mmol of mediator

(hydroquinone). Substrate: H2O2, 176 mmol. Applied potential: �/

0.250 V (vs. Ag/AgCl). For each assay, n�/4.

4.3. Stability of streptavidin carbon-polymer

biocomposites

Due to the biological nature of these biocomposite

electrodes, another important parameter that needs to

be considered is their stability over time. As the

streptavidin is not chemically confined within the

composite electrode, there is a possibility of either loss

of streptavidin stability or the leaching of streptavidin

into either the hybridisation solution or the electrolyte

solution. Such phenomenon has been observed else-

where (Alegret et al., 1996) for other carbon-polymer

biocomposite electrodes. To study the stability of these

biocomposites, four electrodes were prepared and ex-

periments were carried on a weekly basis, over a 12-

week period, during which time the electrodes were

stored in an upright position at 4 8C. All other experi-

mental conditions were maintained constant. In all cases

the electrodes were polished prior to use exposing a fresh

surface with streptavidin that is available for binding,

avoiding the problems derived from breaking the

streptavidin/biotin bond by exposing it to high tempera-

tures or extreme conditions which may adversely affect

the stability of the biocomposite (Lo et al., 1999).As the concern here is the availability of streptavidin

for binding, it was decided to investigate the stability

using biotin�/HRP conjugate, where electrochemical

detection was carried out as described previously in

Section 3.3. The problem of non-specific adsorption

needs to be addressed. Therefore, all experiments were

done in parallel using a blank carbon composite

electrode. Fig. 4 shows that the contribution of non-

specific adsorption of the biotin�/HRP conjugate results

in a current response that is relatively consistent with

time and less than 150 nA in all cases. The results

obtained for the streptavidin biocomposite are also

shown in Fig. 4. After subtracting the response asso-

ciated with non-specific adsorption, these results show a

Fig. 4. Stability of the streptavidin carbon-polymer biocomposite

electrode, time (weeks) vs. current (nA). Non-specific adsorption of

biotin�/HRP on a carbon-polymer composite (') and current

response for streptavidin carbon-polymer biocomposite (j). For each

assay, n�/4. Other experimental details as in Fig. 3.

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175 171

variation over time ranging from a maximum value of

800 nA for week 1 to a minimum value of 185 nA at

week 11. Fig. 4 shows the mean value of a batch of four

electrodes, with a RSD of 15%. However, it is importantto remember that there is no uniform dispersion of

streptavidin within the carbon composite. Hence, one

electrode may expose more streptavidin than the other

and give rise to the large experimental error observed.

While, taking into account the variation of available

streptavidin binding sites, a general decrease in current

with time is observed for the freshly polished biocom-

posite surface suggesting denaturation of streptavidinwithin the biocomposite over long periods of time.

Over the 12-week period these electrodes were stu-

died, a quantitative result was obtained. However, if the

depth of the electrode was extended to 6 mm instead of 3

mm mentioned in Section 2.3 this would allow a greater

number of surface regeneration polishes, therefore,

allowing a greater number of uses.

4.4. Non-specific adsorption evaluation. Study of the

mechanism of the one step immobilisation/hybridisation

procedure

As with the more traditional enzyme immunoassays

the problem of non-specific adsorption continues to be

an issue in the development of genosensors based on

enzyme labelling. It is also necessary to study the one

step immobilisation/hybridisation mechanism. Toachieve this, a capture system with peculiar DNA

sequences (see Section 2.1) was chosen: the digoxigenin

modified probe (responsible for the analytical response)

matching with the target-DNA, but also with the

biotinylated capture probe.

Four assays were done for both evaluations. As

explained in Section 3.5, the first of these assays (i)

was done with an unmodified carbon composite elec-trode (containing no streptavidin), using all three

oligonucleotide components. Assays (ii), (iii) and (iv)

were carried out with streptavidin carbon-polymer

biocomposite electrodes. In assay (iv) (the genosensor

in capture format), all three oligonucleotides compo-

nents were used. In assays (ii) and (iii) the same

procedure as above was followed, but omitting (ii)

target-DNA, and (iii) biotinylated capture probe. Inall cases (assays i, ii, iii and iv) four electrodes were used,

and the same polishing procedure described in Section

2.3 was followed. Results are shown in Fig. 5. Fig. 5A

show the net amperometric signals while Fig. 5B shows

the signal percentage of assays (i), (ii) and (iii) with

respect to the genosensor signal in capture format (assay

iv).

It is useful to compare the results of assays (i), (iii) and(iv) to evaluate non-specific adsorption. By comparing

the results obtained in assays (i) (unmodified carbon-

polymer composite) and (iv) (the genosensor showing

the electrochemical response associated with the strep-

tavidin carbon-polymer biocomposite), the importance

of streptavidin in the carbon composite electrode is

evident from the resulting enhancement of the electro-chemical response. As seen in Fig. 5B, the contribution

of non-specific adsorption to the overall analytical

response was 8.95%.

The difference between the results obtained in assay

(i) and (iii) could be associated with non-specific

adsorption due to non-specific interactions with strep-

tavidin in the biocomposite (perhaps due to hydropho-

bic forces between streptavidin and anti-Dig-HRP). Asit is shown in Fig. 5B, the contribution of non-specific

adsorption onto streptavidin carbon-polymer biocom-

posite to the overall analytical response was 13.54%.

Finally, results in Fig. 5A (i) and (iii) show that such

non-specific adsorption does indeed occur. However, if

this response is compared with that of genosensor in

capture format (iv) it may be said that the overall

contribution is low with respect to the proposed assayand well within experimental error. For assay (i) this

error was 8.95% and for assay (iii) this error is 9.45%.

The comparison of results from assays (ii) and (iv)

with those of (i) and (iii) is useful to clarify the one step

immobilisation/hybridisation step mechanism.

Assay (ii) uses a streptavidin carbon-polymer bio-

composite, but in this instance the target-DNA is

omitted. This assay will indicate the contribution ofhybridisation between the digoxigenin modified probe

and the biotinylated capture probe, as the sequences are

complementary. As shown in Fig. 5B, the contribution

of the assay (ii) to the overall analytical response

obtained with the genosensor in capture format was

higher (18.45%) than the contribution of non-specific

adsorption (evaluated in assays (i) and (iii)). These

results indicate that hybridisation between the digox-igenin modified probe and the biotinylated capture

probe occurs to a lesser extent or that enzyme labelling

is favoured in assay (iv) with respect to assay (ii). In

assay (iv), the 50 bases of the target would act as a

spacer arm, exposing the digoxigenin to the solution to a

larger extent than in assay (ii), conferring flexibility and

favouring the reaction with the antibody anti-Dig-HRP.

Fig. 6 explains the possible reaction mechanism forassays (ii) and (iv).

In assay (ii), the biotinylated probe has two possibi-

lities (a) immobilisation, i.e. reaction with the biotin�/

streptavidin in the heterogeneous phase or (b) hybridi-

sation with the digoxigenin modified probe in solution.

Since the reaction biotin�/streptavidin is extremely fast

(Gonzalez et al., 1999), (a) is more likely to occur than

(b). As seen in Fig. 6, the hybridisation mechanismexpected between biotin-(dA)20 and digoxigenin-(dT)20

is in a heterogeneous phase (c) and not in solution (b).

Comparing the results shown in Fig. 5B it is evident that

hybridisation occurs in a heterogeneous phase (c)

Fig. 5. Evaluation of the non-specific adsorption. (A) (i) carbon-polymer composite (which contains no streptavidin), (ii) streptavidin carbon-

polymer biocomposite omitting target-DNA, (iii) streptavidin carbon-polymer biocomposite omitting biotinylated capture probe and (iv) genosensor

in capture format based on streptavidin carbon-polymer biocomposite. (B) Signal percentage of assay (i), (ii) and (iii) with respect to the genosensor

in capture format (assay iv). For each assay, n�/4. Other experimental details as in Fig. 3.

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175172

between biotin-(dA)20 and digoxigenin-(dT)20, repre-

senting 18.45% of the overall analytical response.

In assay (iv) (as shown in Fig. 6), the biotinylated

probe has three possible courses: immobilisation (a), i.e.

the biotin�/streptavidin reaction in heterogeneous phase

and hybridisation with the digoxigenin modified probe

(b) or with target-DNA (e), both in solution. As seen for

assay (ii), it is expected that the favoured reaction is

immobilisation implied by the biotin�/streptavidin reac-

tion. In this fashion the biotinylated probe would be less

available to hybridise in solution with the DNA-target

and with the digoxigenin-dT(20). Among the possible

hybridisation mechanisms, c, d and f (in Fig. 6, assay iv),

the hybridisation in solution (d) between the DNA

target and the digoxigenin-(dT)20 probe, is more likely

than the mechanism occurring in an heterogeneous

phase (c) (between biotin-(dA)20 and digoxigenin-

(dT)20), and (f) (between biotin-(dA)20 and the target-

DNA).

Fig. 6. Schematic representation of the reaction mechanism for the one step

carbon-polymer biocomposite omitting the target-DNA and for assay (iv) wit

polymer biocomposite. See the text for details.

Finally, hybridisation (g) occurs between the functio-

nalised duplex with digoxigenin and the biotinylated

probe immobilised on the composite.The proposed mechanism for assay (iv) is like a

pseudo two step procedure, occurring in one step. First,

immobilisation happens, followed by hybridisation. It is

in agreement with results shown in Fig. 7A and B for a

two step procedure assay, as explained above. The

amperometric signal produced in these assays was of

the same order confirming the mechanism proposed in

Fig. 6.

4.5. Optimisation of genosensor response

The novelty of this approach is in part attributed to

the simplicity of its design. However, combining the

hybridisation and the immobilisation of DNA in one

analytical step reduces the number of steps within theprotocol and, hence, reduces experimental time. There-

immobilisation/hybridisation procedure, for assay (ii) with streptavidin

h the genosensor in the capture format based on a streptavidin carbon-

Fig. 7. Experimental results for the optimisation of the streptavidin carbon-polymer genosensor, time (min) vs. current (nA). Two step procedures:

(A) variation of the immobilisation time with a constant hybridisation time of 60 min and (B) variation of the hybridisation time with a constant

immobilisation time of 60 min. And (C) variation of time for the one step hybridisation/immobilisation procedure. For each assay, n�/4. Other

experimental details as in Fig. 3.

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175 173

fore, as part of the optimisation procedure, it isimportant to ensure that by combining the immobilisa-

tion and hybridisation steps into a one step procedure

does not result in a compromise between sensitivity and

experimental time. For this reason, the following

optimisation experiments were carried out. In all

experiments, four different electrodes were used and

the electrode surface was renewed before each assay.

The first two assays used a two step procedure, while thefinal assay used a one step procedure. In Fig. 7A�/C the

average of the electrochemical response of these four

electrodes is shown and the bars show the standard

deviation of the replicates.

4.5.1. Immobilisation as the limiting step

A variation of the immobilisation time between 10and 120 min at regular intervals of 10 min was chosen.

For subsequent hybridisation, a time of 60 min was

chosen in agreement with literature (Gingeras et al.,

1987). Results of this study are presented in Fig. 7A

showing a clear relationship between immobilisation

time and current response, with a current of 790 nA

being observed at 60 min. However, for immobilisation

times greater than 60 min a current plateau is observed,suggesting a saturation of all available streptavidin

binding sites. Error analysis shows that in all cases

these electrodes show good reproducibility.

4.5.2. Hybridisation as the limiting step

In this instance, a constant immobilisation time of 60

min was used, as seen in Fig. 7A. Again, a variation ofthe hybridisation time between 10 and 120 min at

regular increments of 10 min was used. Results can be

seen in Fig. 7B. In agreement with the previous study it

is seen that experimental time and current are clearly

related. However, unlike the previous experiment that

showed a current plateau at 60 min, this time current

increases proportionally up to 80 min. Additionally, the

current plateau for this experiment is 10% lower thanthe one observed previously.

Results presented in Fig. 7A and B, suggest that the

hybridisation procedure is the limiting step (in agree-

ment with the proposed mechanism shown in Fig. 6).This result is not surprising considering the strong

binding affinity of biotin to streptavidin (Gonzalez et

al., 1999), with a known association constant in the

range of 1015/M. For such a strong binding reaction, one

would not expect the immobilisation of the capture

probe to be the limiting step in the overall procedure.

The experiments described establish the step that

limits the overall electrochemical response. They alsohelp estimate the optimum time for the one step

immobilisation/hybridisation procedure. However, it is

also possible that combining the immobilisation/hybri-

disation in one step will result in a loss of sensitivity.

Therefore a study of the experimental time vs. the

current response was done and results are shown in Fig.

7C. In agreement with Fig. 7A and B it is observed that

current increases with an increase of the immersion timefor times below 60 min. However, for times over 60 min

a current plateau is noticed. These results suggest that

all the available binding sites of the streptavidin

molecules exposed at the electrode surface have been

occupied within this time*/in agreement with results

shown in Fig. 7A.

Fig. 7A�/C shows that separation of hybridisation and

immobilisation in two individual steps does not increasethe amount of HRP available for detection. It should be

noted that optimisation of the enzyme labelling time of

the binding of anti-Dig-HRP to Dig-DNA has been

previously reported (Pividori, 2002) and the time of 40

min was chosen for the present study.

Based on the data presented in Fig. 7 it was decided to

work with the one step hybridisation procedure, as it

does not result in a loss of sensitivity due to a shortertime.

5. Conclusion

The proposed genosensor design has proven to besuccessful in using a simple bulk modification step,

hence, overcoming the complicated pre-treatment steps

associated with other genosensor designs. Also, the use

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175174

of a one step immobilisation and hybridisation proce-

dure reduces the cost of the experiment in addition to a

reduction in experimental time. Such a genosensor

design results a novel approach to DNA immobilisationbased on the well known biotin�/streptavidin interac-

tion. Attaching an enzyme-labelled marker to the DNA

duplex formed on the streptavidin carbon-polymer

biocomposite permits the detection of very small quan-

tities of DNA. An additional novel aspect of this work is

the capability of surface regeneration of the biocompo-

site electrodes allowing repeated analyses with the same

electrode without need to denature the dsDNA on theelectrode surface. Stability studies conducted demon-

strate the capability of the same electrode to be used for

a 12-week period.

The new material developed can be seen as a universal

electrochemical platform since it can be used to detect

any analyte that can be coupled to the biotin�/strepta-

vidin reaction as is the case of immunoassays in addition

to its use in DNA analysisFuture work will be focused on the possible applica-

tion of such method to screen printed electrodes,

allowing the development of disposable electrodes cap-

able of a quantitative DNA detection, in a simple, low

cost and easy to manipulate manner.

Acknowledgements

We thank the Ministerio de Educacion y Ciencia de

Madrid (Project BIO2000-0681-C02-01) for financial

assistance and Edna Williams thanks the Departamento

de Educacion y Ciencia and Higher Education Author-

ity for additional financial support.

References

Alegret, S., 1996. Rigid carbon-polymer biocomposites for electro-

chemical sensing. Rev. Analyst 121, 1751�/1758.

Alegret, S., Alonso, J., Bartrolı, J., Cespedes, F., Martınez-Fabregas,

E., del Valle, M., 1996. Amperometric biosensors based on bulk-

modified epoxy graphite biocomposites. Sensor Mater. 8 (3), 147�/

153.

Armistead, P.M., Thorp, H.H., 2000. Modification of indium tin oxide

electrodes with nucleic acids: detection of attomole quantities of

immobilised DNA by electrocatalysis. Anal. Chem. 72, 3764�/3770.

Armistead, P.H., Thorp, H.H., 2001. Oxidation kinetics of guanine in

DNA molecules adsorbed onto indium tin oxide electrodes. Anal.

Chem. 73, 558�/564.

Alfonta, L., Singh, A.K., Wollner, I., 2001. Liposomes labelled with

biotin and horseradish peroxidase: a probe for enhanced amplifica-

tion of antigen�/antibody or oligonucleotide*/DNA sensing pro-

cesses by precipitation of an insoluble product on electrodes. Anal.

Chem. 73, 91�/102.

Burrows, C.J., Muller, J.G., 1998. Metallodrug complexes that

mediate DNA and lipid damage via sulfide autoxidation: cop-

per(II) famotidine and iron(III) bis(salicylglycine). Inorg. Chim.

Acta 275�/276, 314�/319.

Chen, X., Ruan, C., Kong, J., Deng, J., 2000. Characterisation of

direct electron transfer and bioelectrocatalysis of horseradish

peroxidase in DNA films at pyrolytic graphite electrodes. Anal.

Chem. Acta 412, 89�/98.

Chiti, G., Marrazza, G., Mascini, M., 2001. Electrochemical DNA

biosensor for enviromental monitoring. Anal. Chim. Acta 427,

155�/164.

Cespedes, F., Martınez-Fabregas, E., Alegret, S., 1996. New materials

for electrochemical sensing. I. Rigid conducting composites. TrAC

15, 296�/303.

Erkkila, K.E., Odom, D.T., Barton, J.K., 1999. Recognition and

reaction of metallointercallators with DNA. Chem. Rev. 99, 2777�/

2795.

Gingeras, T.R., Kwoh, D.Y., Davis, G.R., 1987. Hybridization

properties of immobilized nucleic acids. Nucleic Acids Res. 15,

5373�/5390.

Gonzalez, M., Argarana, C.E., Fidelio, G.D., 1999. Extremely high

thermal stability of streptavidin and avidin upon biotin binding.

Biomol. Eng. 19, 67�/72.

Hashimoto, K., Ito, K., Ishimori, Y., 1994. Novel DNA sensors for

electrochemical gene detection. Anal. Chim. Acta 286, 219�/224.

Hashimoto, K., Ito, K., Ishimori, Y., 1998. Sequence-specific gene

detection with a gold electrode modified with DNA probes and an

electrochemically active dye. Anal. Chem. 66, 3830�/3833.

Herne, T.M., Tarlov, M.J., 1997. Characterization of DNA probes

immobilized onto gold surfaces. J. Am. Chem. Soc. 119, 8916�/

8920.

Kizek, R., Ludek, H., Frojta, M., Palecek, E., 2002. Determination of

nanogramquantities of osmium labelled single stranded DNA by

differential pulse stripping voltammetry. Bioelectrochemistry 55,

119�/121.

Kricka, L.J., 1992. Non-isotopic DNA Probe Techniques. Academic

Press, Inc, San Diego.

Lee, T.-Y., Shim, Y.-B., 2001. Direct DNA hybridisation detection

based on the oligonucleotide*/funtionalised conductive polymer.

Anal. Chem. 73, 5629�/5632.

Li, L., Wang, M., Dong, S., Wang, E., 1997. Electrochemical

investigation of DNA adsorbed on conducting polymer modified

electrodes. Anal. Sci. 13, 305�/310.

Lo, Y.-S., Huefner, N.D., Chan, W.S., Stevens, F., Harris, J.H.M.,

Beebe, T.P., 1999. Specific interactions between biotin and avidin

studied by atomic force spectroscopy using the Poisson statistical

analysis method. Langmuir 15, 1373�/1982.

Lumley-Woodyear, T., Campbell, C.N., Freeman, E., Freeman, A.,

Georgiou, G., Heller, A., 1999. Rapid amperometric verification of

PCR amplification of DNA. Anal. Chem. 71, 535�/538.

Marrazza, G., Chianella, I., Mascini, M., 1999. Disposable DNA

electrochemical biosensors for environmental monitoring. Anal.

Chim. Acta 387, 297�/307.

Millan, K.M., Mikkelsen, S.R., 1993. Sequence-selective biosensor for

DNA based on an electroactive hybridisation indicators. Anal.

Chem. 65, 2317�/2323.

Millan, K.M., Spurmanis, A.L., Mikkelsen, A.L., 1992. Covalent

immobilisation of DNA onto glassy carbon electrodes. Electro-

analysis 4, 929�/932.

Nuzzo, R.G., Allara, D.L., 1983. Adsorption of bifunctional organic

disulphides on gold surfaces. J. Am. Chem. Soc. 105, 4481�/4483.

Palecek, E., 1958. Oscillographic polarograhpy of nucleic acids and

their building blocks. Naturwiss 45, 186�/187.

Palecek, E., 1960. Oscillographic polarography of highly polymerized

deoxyribonucleic acid. Nature 188, 656�/657.

Palecek, E., Fojta, M., 2001. DNA hybridization and damage. Anal.

Chem., 75A�/83A.

Palecek, E., Kizek, R., Havran, L., Billova, S., Fojta, M., 2002.

Electrochemical enzyme linked immunoassay in a DNA hybridisa-

tion sensor. Anal. Chim. Acta 469, 73�/83.

E. Williams et al. / Biosensors and Bioelectronics 19 (2003) 165�/175 175

Pantano, P., Morton, T.H., Kuhr, W.G., 1991. Enzyme-modified

carbon-fiber microelectrodes with millisecond response times. J.

Am. Chem. Soc. 113, 1832�/1833.

Pividori, M.I., 2002. Ph.D. Thesis. Autonomous University of

Barcelona, Spain.

Pividori, M.I., Merkoci, A., Alegret, S., 2000. Electrochemical

genosensor design: immobilisation of oligonucleotides onto trans-

ducer surfaces and detection methods. Biosen. Bioelectron. 15,

291�/303.

Pividori, M.I., Merkoci, A., Alegret, S., 2001a. Classical dot blot

format implemented as an amperometric hybridisation genosensor.

Biosen. Bioelectron. 16, 1133�/1142.

Pividori, M.I., Merkoci, A., Alegret, S., 2001b. Dot-blot amperometric

genosensor for detecting a novel determinant of b-lactamase

resistance in Staphylococcus aureus . Analyst 126, 1551�/1557.

Radi, A., 1999. Volumetric study of Khellin at DNA-coated carbon

paste electrodes. Anal. Chim. Acta 386, 63�/68.

Rajksi, S.R., Jackson, B.A., Barton, J.K., 2000. DNA repair models

for damage and mismatch recognition. Mutat. Res. 447, 49�/72.

Shamansky, L.M., Brandon-Davis, C., Stuart, J.K., Kuhr, W.G.,

2001. Immobilisation and detection of DNA on microfludic chips.

Talanta 55, 909�/918.

Santandreu, M., Cespedes, F., Alegret, S., Martınez-Fabregas, E.,

1997. Amperometric immunosensors based on rigid conducting

immunocomposites. Anal. Chem. 69, 2080�/2085.

Takenaka, S., Uto, Y., Kondo, H., Ihara, T., Takagi, M., 1994.

Electrochemically active DNA probes. Detection of target DNA

sequences at femtomole level by High Performance Liquid

Chromatography with electrochemical detection. Anal. Biochem.

218, 436�/443.

Turner, S.J., Lewis, G.D., Bellamy, A.R., 1997. Detection of sewage-

derived Escherichia coli in a rural stream using multiplex PCR and

automated DNA detection. Water Sci. Technol. 35, 337�/342.

Uto, Y., Kondo, H., Abe, M., Suzuki, T., Takenaka, S., 1997.

Electrochemical analysis of DNA amplified by the polymerase

chain reaction with a ferrocenylated oligonucleotide. Anal. Bio-

chem. 250, 122�/124.

Wang, J., Cai, X., Jonsson, C., 1996. Adsorptive stripping potentio-

metry of DNA at electrochemically pre-treated carbon paste

electrodes. Electroanalysis 8, 20�/24.

Wang, J., Cai, X., Fernandes, J.R., Grant, D.H., Ozsoz, M., 1997a.

Electrochemical measurements of oligonucleotides in the presence

of chromosomal DNA using membrane-covered carbon electrodes.

Anal. Chem. 69, 4056�/4059.

Wang, J., Cai, X., Rivas, G., Shiraishi, H., Dontha, N., 1997b. Nucleic

acid immobilisation recognition and detection at chronoampero-

metric DNA chips. Biosen. Bioelectron. 12 (7), 587�/599.

Wang, J., Fernandes, J.R., Kubota, L.T., 1998. Polishable and

renewable DNA hybridisation biosensors. Anal. Chem. 70,

3699�/3702.

Wang, J., Kawde, A.-N., 2001. Pencil based renewable biosensors for

label free electrochemical detection of DNA hybridisation. Anal.

Chim. Acta 431, 219�/224.

Xu, C., Cai, H., He, P., 2001. Electrochemical detection of sequence

specific DNA using a DNA probe labelled with aminoferrocene

and chitosan modified electrode immobilised with ssDNA. Analyst

126, 62�/65.