Theory and practice of enzyme bioaffinity electrodes. Direct electrochemical product detection

Published on Web 05/20/2008

Theory and Practice of Enzyme Bioafﬁnity Electrodes. Direct

Electrochemical Product Detection

Benoˆıt Limoges,*^,†Damien Marchal,^†François Mavré,^†Jean-Michel Savéant,*^,†and

Bernd Schöllhorn^‡

Laboratoire d’Electrochimie Moléculaire, UniVersité Paris Diderot, UMR CNRS 7591,

2 place Jussieu, 75251 Paris Cedex 05, France, and Département de Chimie, Ecole Normale

Supérieure, UMR CNRS 8640-PASTEUR, 24 rue Lhomond, 75231 Paris Cedex 05, France

Received November 22, 2007; E-mail: limoges@univ-paris-diderot.fr; saveant@univ-paris-diderot.fr

Abstract: The use of enzyme labeling techniques to convert biorecognition events into high sensitivity

electrochemical signals may follow two different strategies. One, in which the current is the electrocatalytic

response of a redox couple serving as cosubstrate to a redox enzyme label and another that consists in

the detection of an electrochemically active product of the enzyme label. The theoretical relationships that

link, in the latter case, the electrochemical current response to the amount of recognized labeled target

analyte are established for steady-state diffusion-convection chronoamperometric regimes. Two governing

parameters thus emerge. One measures the Michaelis–Menten competition in the enzyme kinetics. The

other characterizes the competition between the enzymatic kinetics and the diffusion of the substrate. The

electrochemical response is ﬁnally related to the labeled target analyte concentration in solution through

the recognition isotherm. The direct electrochemical product detection thus provides a route to the

characteristics of the recognition isotherm, which serves as a calibration curve in analytical applications.

The establishment of further theoretical relationships allows one to surmise the increase in sensitivity that

may be obtained by using cyclic voltammetry instead of steady-state chronoamperometry in standard

electrochemical cells or by accumulation of the enzyme–product in cells of small volume/surface ratios.

The theoretical predictions are tested with the example of the avidin–biotin recognition process in a system

that involves alkaline phosphatase as enzyme label and 4-amino-2,6-dichlorophenyl phosphate as substrate,

generating 4-amino-2,6-dichlorophenol as electrochemically active product. The advantages of the dichloro-

substitution are discussed. The theoretical analysis is a requisite for a rational and realistic discussion of

the analytical performances of the steady-state chronoamperometric and cyclic voltammetric approaches.

These are shown to compare favorably with the best heterogeneous bioafﬁnity assays so far reported.

Introduction

both as an immobilization platform and as a detector to reveal,

with the help of an enzyme label, the speciﬁc bimolecular

Enzyme-ampliﬁed bioafﬁnity electrodes, based on a recep-

tor–ligand binding, for example, immuno- and DNA-sensors,

have attracted considerable attention as biosensors in recent

years, and have emerged as viable alternatives to conventional

spectrophotometric enzyme afﬁnity assays.^1–4They can reach

low-level detection with a high selectivity in small volumes (not

limited, as with optical methods, by the Beer–Lambert law).

They are also robust, economical to mass produce, and can be

easily miniaturized to circuit board level with low power

requirements, being thus good candidates for the development

of ﬁeld-portable analytical instrumentations and integration into

multiarrays, microﬂuidic chips, and other microsensing devices

for multianalyte detection. Assays by means of enzyme-

ampliﬁed bioafﬁnity electrodes are a particular class of hetero-

geneous enzyme bioafﬁnity assays in which the electrode serves

recognition between the immobilized ligand (e.g., antibody,

oligonucleotide, carbohydrate, biotin) and a target receptor (e.g.,

antigen, DNA fragment, protein, avidin) present in solution. The

enzyme label not only effects the transduction of the biomo-

lecular recognition event into an electrochemical signal but also

allows its ampliﬁcation according to the catalytic activity of

the enzyme, leading to the possibility of developing highly

sensitive afﬁnity sensors. The fact that recognition occurs at

the electrode surface entails that the enzymatic reaction takes

place in its close vicinity leading to a fast setting up of the

electrochemical response. It also opens a route to an array of

electrodes for multianalyte analysis, taking advantage of the

spatial discrimination of the enzyme-ampliﬁed electrochemical

response.⁵Enzyme labeling techniques converting a biorecog-

nition event into an electrochemical signal may be applied in

two ways. In one of these, the current is the electrocatalytic

response of a redox couple serving as cosubstrate to a redox

enzyme label (Scheme 1a).^6–11The other consists in the

^†Université Paris Diderot.

^‡Ecole Normale Supérieure.

(1) Díaz-González, M.; González-García, M. B.; Costa-García, A. Elec-

troanalysis 2005, 17, 1901–18.

(2) Skladal, P. Electroanalysis 1997, 9, 737–45.

(3) Killard, A. J.; Smyth, M. R. Anal. Lett. 2000, 33, 1451–65.

(4) Wang, J. Biosens. Bioelectron 2006, 21, 1887–92.

(5) Wilson, M. S.; Nie, W. Anal. Chem. 2006, 78, 2507–13.

9

10.1021/ja7102845 CCC: $40.75

2008 American Chemical Society

J. AM. CHEM. SOC. 2008, 130, 7259–7275 7259

A R T I C L E S

Limoges et al.

Scheme 1

Despite the large number of enzyme-ampliﬁed bioafﬁnity

electrodes described in the literature, little effort has been made

so far to establish rigorously the relationships that link the

electrochemical responses to the surface and bulk concentrations

of the enzyme-labeled target analyte, which form the necessary

basis for determining the recognition binding characteristics as

well as for a realistic evaluation of the analytical performances

(sensitivity, detection limit, linear dynamic range of the analyte

concentration) of such electrodes. The few exceptions to this

lack of theoretical analyses have concerned the redox enzyme-

labeled systems^42a,band not the systems based on the electro-

chemical detection of the enzyme label product.

The aim of the present two companion papers is to ﬁll this

theoretical gap. The ﬁrst paper is devoted to the various modes

of direct electrochemical detection. In the second,⁴³we discuss

chemical and enzymatic means of amplifying the electrochemi-

cal response.

Concerning direct electrochemical detection, we ﬁrst establish

the relationships that link the electrochemical current response

to the amount of recognized labeled target analyte for a steady-

state diffusion-convection chronoamperometric regime. The

electrochemical response is then related to the labeled target

analyte concentration in solution through the recognition

isotherm, thus offering a way to determine its characteristics

and leading to a rational estimation of the analytical perfor-

mances. The establishment of further theoretical relationships

allows the estimation of the increase in sensitivity that may be

obtained by using cyclic voltammetry instead of steady-state

diffusion-natural convection chronoamperometry in a standard

electrochemical cell or by accumulation of the enzyme–product

in a system of very low volume/surface ratios.

The theoretical analysis is then illustrated with the avidin–

biotin recognition process in a system that involves alkaline

phosphatase (AP) as enzyme label and 4-amino-2,6-dichlo-

rophenyl phosphate as substrate, generating 4-amino-2,6-dichlo-

rophenol as the electrochemically active product. More precisely,

the system is as depicted in Scheme 2, showing the binding of

alkaline phosphatase-conjugated neutravidin (N-AP)⁴⁴to a

monolayer of biotinylated immunoglobulin (b-IgG) irreversibly

adsorbed on the surface of the electrode,⁴⁵followed by the

detection of an electrochemically active product of the enzyme

label (Scheme 1b).^5,12–41

(6) Campbell, C. N.; De Lumley-Woodyear, T.; Heller, A. Fresenius J.

Anal. Chem. 1999, 364, 165–9.

(7) Rishpon, J.; Ivnitski, D. Biosens. Bioelectron. 1997, 12, 195–204.

(8) Caruana, D. J.; Heller, A. J. Am. Chem. Soc. 1999, 121, 769–74.

(9) Kavanagh, P.; Leech, D. Anal. Chem. 2006, 78, 2710–6.

(10) Kim, H. H.; Zhang, Y.; Heller, A. Anal. Chem. 2004, 76, 2411–4.

(11) De Lumley-Woodyear, T.; Campbell, C. N.; Heller, A. J. Am. Chem.

Soc. 1996, 118, 5504–5.

(12) Azek, F.; Grossiord, C.; Joannes, M.; Limoges, B.; Brossier, P. Anal.

Biochem. 2000, 284, 107–13.

(13) Kojima, K.; Hiratsuka, A.; Suzuki, H.; Yano, K.; Ikebukuro, K.;

Karube, I. Anal. Chem. 2003, 75, 1116–22.

(14) Volpe, G.; Fares, G.; delli Quadri, F. ; Draisci, R.; Ferretti, G.;

Marchiafava, C.; Moscone, D.; Palleschi, G. Anal. Chim. Acta 2006,

572, 11–6.

(15) Butler, D.; Guilbault, G. G. Sens. Actuators, B 2006, B113, 692–9.

(16) Diaz-Gonzalez, M.; Gonzalez-Garcia, M. B.; Costa-Garcia, A. Sens.

Actuators, B 2006, B113, 1005–11.

(17) Xu, Y. F.; Velasco-Garcia, M.; Mottram, T. T. Biosens. Bioelectron.

2005, 20, 2061–70.

(33) Hernández-Santos, D.; Diaz-González, M.; Gonzalez-García, M. B.;

Costa-García, A. Anal. Chem. 2004, 76, 6887–93.

(34) Abad-Valle, P.; Fernandez-Abedul, M. T.; Costa-Garc´ıa, A. Biosens.

Bioelectron. 2005, 20, 2251–60.

(35) Liu, D.; Perdue, R. K.; Sun, L.; Crooks, R. M. Langmuir 2004, 20,

5905–10.

(36) Nebling, E.; Grunwald, T.; Albers, J.; Schaefer, P.; Hintsche, R. Anal.

Chem. 2004, 76, 689–96.

(37) Patolsky, F.; Lichtenstein, A.; Willner, I. Chem. Eur. J. 2003, 9, 1137–

45.

(38) Patolsky, F.; Lichtenstein, A.; Willner, I. Nat. Biotechnol. 2001, 19,

253–7.

(39) Lucarelli, F.; Marrazza, G.; Mascini, M. Langmuir 2006, 22, 4305–9.

(40) Fojta, M.; Brazdilova, P.; Cahova, K.; Pecinka, P. Electroanalysis 2006,

18, 141–51.

(41) Mittelmann, A. S.; Ron, E. Z.; Rishpon, J. Anal. Chem. 2002, 74,

903–7.

(18) D´ıaz-Gonza´lez, M.; Gonza´lez-Garc´ıa, M. B.; Costa-Garc´ıa, A. Biosens.

Bioelectron 2005, 20, 2035–43.

(19) Wilson, M. S. Anal. Chem. 2005, 77, 1496–502.

(20) Ammida, N. H. S.; Volpe, G.; Draisci, R.; delli Quadri, F.; Palleschi,

L.; Palleschi, G. Analyst 2004, 129, 15–9.

(21) Vetcha, S.; Wilkins, E.; Yates, T. Biosens. Bioelectron. 2002, 17,

901–9.

(22) O’Regan, T. M.; O’Riordan, L. J.; Pravda, M.; O’Sullivan, C. K.;

Guilbault, G. G. Anal. Chim. Acta 2002, 460, 141–50.

(23) Pemberton, R. M.; Hart, J. P.; Stoddard, P.; Foulkes, J. A. Biosens.

Bioelectron. 1999, 14, 495–503.

(24) Schreiber, A.; Feldbrugge, R.; Key, G.; Glatz, J. F. C.; Spener, F.

Biosens. Bioelectron. 1997, 12, 1131–7.

(25) Rosen, I.; Rishpon, J. J. Electroanal. Chem. 1989, 258, 27–39.

(26) Duan, C.; Meyerhoff, M. E. Anal. Chem. 1994, 66, 1369–77.

(27) Aguilar, Z. P.; Vandaveer, W. R.; Fritsch, I. Anal. Chem. 2002, 74,

3321–9.

(42) (a) Gyss, C.; Bourdillon, C. Anal. Chem. 1987, 59, 2350–5. (b)

Bourdillon, C.; Demaille, C.; Moiroux, J.; Savéant, J-M. J. Am. Chem.

Soc. 1999, 121, 2401–8.

(28) Elsholz, B.; Woerl, R.; Blohm, L.; Albers, J.; Feucht, H.; Grunwald,

T.; Juergen, B.; Schweder, T.; Hintsche, R. Anal. Chem. 2006, 78,

4794–802.

(43) Limoges, B.; Mavré, F.; Marchal, D.; Savéant, J.-M. J. Am. Chem.

Soc. 2008, 130, 7276–7285.

(29) Bagel, O.; Degrand, C.; Limoges, B.; Joannes, M.; Azek, F.; Brossier,

P. Electroanalysis 2000, 12, 1447–52.

(44) Neutravidin, a deglycosylated form of avidin, was preferred to avidin

itself because it minimizes nonspeciﬁc absorption effects.

(45) With the aim to ﬁll the spaces remaining vacant on the electrode after

adsorption of b-IgG and thus to minimize nonspeciﬁc adsorption during

the recognition step, the electrode is exposed to bovine serum albumin

(BSA) (see Scheme 2).

(30) Aguilar, Z. P.; Fritsch, I. Anal. Chem. 2003, 75, 3890–7.

(31) Aguilar, Z. P. Anal. Chem. 2006, 78, 1122–9.

(32) Carpini, G.; Lucarelli, F.; Marrazza, G.; Mascini, M. Biosens.

Bioelectron. 2004, 20, 167–75.

9

7260 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

Enzyme Bioafﬁnity Electrodes. Direct Detection

A R T I C L E S

Scheme 2

Scheme 3

behaves more satisfactorily in this respect. We have thus

synthesized the parent AP substrate, 4-amino-2,6-dichlorophenyl

phosphate (2,6-DCPAPP or S), and used it to illustrate the

theoretical analysis, determine the afﬁnity binding constant, and

characterize the electroanalytical performances of the enzyme

afﬁnity electrodes toward the quantitative detection of AP-

conjugated avidin, as depicted in Scheme 2. The AP label

speciﬁcally bonded to the electrode was determined by both

amperometry and cyclic voltammetry through the electrochemi-

cal oxidation of the enzyme generated 2,6-DCPAP product (P)

into a 2,6-dichloro-quinonimine (2,6-DCQI or Q) at potentials

that do not interfere with the irreversible oxidation of the

substrate S. The kinetic parameters of the enzymatic dephos-

phorylation were determined in homogeneous solution and were

used to estimate, from the steady-state amperometric response

or the voltammetric peak current, the amount of AP bonded

onto the electrode and its maximal coverage. To check the

assumption that the kinetic and activity of AP immobilized onto

the electrode is the same as in a homogeneous solution, the AP

coverage at saturation was also determined by an indirect

method. Addressing the important problem of the blank current

response in the absence of AP led us to examine the purity and

stability of both the substrate 2,6-DCPAPP and the product 2,6-

DCPAP.

On the basis of the parameters that emerge from the

theoretical analysis the analytical performances are then dis-

cussed so as to relate their progressive improvement to the

degree of sophistication of the technique (from steady-state

chronoamperometry to cyclic voltammetry) and of the cell (from

conventional cells to low volume-to-surface ratio cells) and

compared with the performances of nonelectrochemical ap-

proaches. These parameters also help to optomize the choice

of the enzyme by means of the kinetic characteristics they

contain.

electrochemical detection of the activity of the AP label

speciﬁcally attached on the electrode surface.

The reason for selecting alkaline phosphatase was its ability

to rapidly convert a substrate (S) into an electrochemically

detectable product (P). This is also why AP has been widely

used as an enzyme label in numerous amperometric enzyme

immunosensors^5,14–27or DNA-sensors^28–41and is at the origin

of some of the most sensitive enzyme-ampliﬁed biosensing

applications. AP catalyzes the hydrolysis of aromatic monoester

phosphates to give phenolic derivatives and an inorganic

phosphate. A broad range of aromatic monoester phosphate

substrate/phenolic product couples has thus been proposed for

the electrochemical detection of AP and consequently applied

to electrochemical enzyme afﬁnity assays.

Among these substrate/product couples, the p-aminophenyl

phosphate (PAPP)/p-aminophenol (PAP) couple has been found

to be the most advantageous because of the reversible redox

behavior of the PAP product at a bare electrode, taking place

at a low potential (formal potentials: 0.1 V vs SCE at pH 9),

and also because of its low propensity to foul the electrode

surface compared to other phenolic products.^46–49However, as

shown in the present report, the electrochemical behavior of

PAP is strongly affected by the presence of macrobiomolecules

on the electrode surface, leading to a poorly reversible cyclic

voltammetric wave endowed with an apparent heterogeneous

electron transfer rate signiﬁcantly lower than at a bare electrode.

We have found that 4-amino-2,6-dichlorophenol (2,6-DCPAP)

Results and Discussion

1. Theoretical Relationships. 1.1. Chronoamperometric Re-

sponse under Steady State Natural Convection/Diffusion Condi-

tions. We consider that, in the absence of stirring, a steady-

state current is rapidly established after applying the potential

on account of the natural convection that takes place in solution.

The diffusion-reaction problem is thus as sketched in Scheme

3. The electrode potential is poised at a value sufﬁciently

positive for the concentration of the enzyme-generated product

P to be equal to zero at the electrode surface. The enzyme

(46) Frew, J. E.; Foulds, N. C.; Wilshere, J. M.; Forrow, N. J.; Green,

M. J. J. Electroanal. Chem. 1989, 266, 309–16.

(47) Wilson, M. S.; Rauh, R. D. Biosens. Bioelectron. 2004, 20, 276–83.

(48) Kreuser, M. P.; O’Sullivan, C. K.; Guilbault, G. G. Anal. Chim. Acta

1999, 393, 95–102.

(49) Tang, H. T.; Lunte, C. G.; Halsall, H. B.; Heineman, W. R. Anal.

Chim. Acta 1988, 214, 187–95.

9

J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7261

A R T I C L E S

Limoges et al.

the enzyme kinetics is assumed to obey the steady state

approximation, and since Γ_ES+ Γ_E) Γ_E⁰

[S]_{x ) 0}

^EK_M+ [S]_{x ) 0}

Γ_ES) Γ⁰

Therefore,

[S]_{x ) 0}

^EK_M+ [S]_{x ) 0}

I

2F

) k₂Γ⁰

The substrate concentration at the electrode surface may be

expressed as a function of the current as

[S]

^{x ) 0}) 1 -

I

2FD_SC_S⁰

δ

C_S⁰

where C⁰_Sis the substrate bulk concentration and δ is the

thickness of the natural convection/diffusion layer. The term

2FD_SC_S⁰

I_S)

(1)

δ

is the plateau current density of the substrate that would be

observed if it were directly oxidized at the electrode. Thus

Figure 1. Variations of the current density, under the form of I/I_k(green)

or I/I_S, (blue), according to eq 3 and 3′, with the two parameters I_k/I_S, and

K_M/C⁰_S(the number in each diagram is the value of log(K_M/C⁰_S))

I

I_S

1 -

I_k

I

)

I_SI_SK_M

I

I_S

+ 1 -

kinetics is assumed to comply with the classical Michaelis–

Menten format, where E designates the free enzyme and ES

the intermediate enzyme–substrate complex. The notations for

the rate constants are indicated in the scheme.

C_S⁰

where besides the substrate diffusion current density, a current

density, I_k, characterizing the enzymatic kinetics

In a ﬁrst step, we look for the relationship that links the

current density, I, to the surface concentration of the recognized

enzyme-labeled analyte, Γ⁰_E. We consider that the concentration

of substrate and/or the volume of solution are large enough for

its consumption in the bulk to be negligible. The current may

then be expressed by

I_k) 2Fk₂Γ⁰_E

(2)

has been introduced

K_MI_k

I

2

I

-

+

^k) 0

0

( )

(

)

I_S

I_SI_SI_S

C_S

I

2F

d[Q]

dx x ) 0

) -D_Q

(

)

or, alternatively, if the current density is normalized toward the

enzyme kinetic current density, I_k, rather than toward the

substrate diffusion current density:

d[S]

dx x ) 0

) k₁[S]_{x ) 0}Γ_E- k _{- 1}Γ_ES

)D_S

(

)

₂I

K_MI_k

^k- 1 +

+

+ 1 ) 0

I

) k₂Γ_ES

C_S⁰

( )

(

)

I_kI_S

I_SI_k

where D is the diffusion coefﬁcient of the subscript species, [ ]

represents the corresponding concentration, and Γ is the surface

concentration of the two forms of the enzyme. The subscript x

) 0 indicates that the concentrations and ﬂuxes are taken at

the electrode or at the enzyme layer. The two corresponding

average planes may be confounded since the thickness of the

enzyme ﬁlm is small as compared to the diffusion layer (the

thicknesses in Scheme 3 are arbitrary). The factor 2 in the ﬁrst

member of the above equation is a stoichiometric factor

expressing the fact that the electrochemical oxidation of the

product consumes two electrons per molecule as in the illustrat-

ing example below and in a number of other cases.

As soon as it is formed at the electrode surface by the

enzymatic reaction, the product P is oxidized into Q owing to

the poising of the electrode potential at a value positive enough

for its concentration to be zero there, thus preventing any

diffusion toward the solution. As indicated in the above equation,

Two competition parameters thus emerge from the above

equations, K_M/C⁰_S, related to the Michaelis–Menten kinetic

competition and I_k/I_S) k₂Γ⁰_Eδ/D_SC⁰_S, which measures the

competition between the substrate diffusion and enzyme kinetics.

Examples of variations of the current density, under the form

of I/I_k, or I/I_S, with the two parameters are shown in Figure 1.

Increasing the bulk substrate concentration decreases each

parameter leading to a limiting situation where the current is

solely governed by a saturated Michaelis–Menten kinetics with

no interference of substrate diffusion:

I ) I_k) 2Fk₂Γ⁰_E

This is the best situation for maximizing the current response

for a given enzyme. If this limiting situation is not reached,

one may use the more complete relationships, which may be

recast as

9

7262 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

Enzyme Bioafﬁnity Electrodes. Direct Detection

A R T I C L E S

cm, it leads for ν ) 0.1 V/s and D ) 4 × 10^-6cm²s^-1to

δ/(DRT/FV)^1/2) 10. Under such conditions the peak current is

proportional to the steady state concentration of P at the

electrode surface, which is related to the various parameters of

the system as follows. Unlike the situation in section 1.1, P is

not oxidized at the electrode and therefore diffuses in the

solution. The steady-state production of P equals the consump-

tion of the substrate S according to⁵⁰

2

K_MI_k

C_S⁰

I_S

K_MI_k

I_k

I_S

1 +

+

-

1 +

+

- 4

C_S⁰

(

)

I_S

ꢀ

I

I_S

)

(3)

2

or

2

I_SK_MI_S

I_S

I_k

1 +

+

-

1 +

+

- 4

0

(

)

I

I_k

I

I_k

^kC_S

ꢀ

I

I_k

d[P]

dx x ) 0

)k₁[S]_{x ) 0}Γ_E- k_-1Γ_ES) k₂Γ_ES

d[S]

dx x ) 0

)

(3′)

-D

) D_S

(

)

(

)

2

(5)

It is noteworthy that even when K_M/C⁰_Sis small as in the left-

hand diagrams in Figure 1, the rate-control passes from

enzymatic kinetics to substrate diffusion upon increasing I_k/I_S.

The transition between one kinetic control and the other is the

sharper the smaller K_M/C⁰_Sis.

Another limiting situation of interest is when I_k/I_Sis small

and K_M/C⁰_Sis not particularly small (a situation that can be

reached for low surface concentration of the enzyme-labeled

analyte, Γ⁰_E). The current density is then given by

As before,

and

[S]_{x ) 0}

^EK_M+ [S]_{x ) 0}

Γ_ES) Γ⁰

[S]_{x ) 0}

^EK_M+ [S]_{x ) 0}

C⁰_S- [S]_{x ) 0}

d[S]

dx x ) 0

D_S

) k₂Γ⁰

) D_S

(

)

δ

C_S⁰

K_M+ C_S⁰

C_S⁰

Thus

I ) I_k

) 2Fk₂Γ⁰_E

(4)

K_M+ C_S⁰

[S]_{x ) 0}

C_S⁰

[S]_{x ) 0}

Once the relationship between the current density and the surface

concentration of the recognized labeled target analyte has been

established, the surface concentration can be related to its bulk

concentration during the afﬁnity-binding step. The recognition

isotherm (for example, but not necessarily,⁴³a Langmuir isotherm)

may thus be obtained, leading to the determination of the binding

equilibrium constant. Relating ﬁnally the current density to the

solution concentration of the labeled target analyte will allow the

estimation of the analytical performances of the method.

1.2. Cyclic Voltammetry at High Volume-to-Surface Ratios. An

alternative approach consists in letting the enzyme produce the

product P under steady-state natural convection/diffusion condi-

tions and then to sample the production of P by means of cyclic

voltammetry as sketched in Scheme 4, under conditions where

the scan rate is large enough for the thickness cyclic voltam-

metry diffusion layer, (DRT/FV)^1/2(D is the value of the

diffusion coefﬁcient of P, which is assume to be approximately

equal to the diffusion coefﬁcient of Q), to be small compared

to that of the steady state natural convection/diffusion layer, δ.

A scan rate not larger than 0.1 V/s is usually sufﬁcient for this

condition to be fulﬁlled. For example, if δ is as small as 0.01

I_k

[S]_{x ) 0}

C_S⁰

) 1 -

I_SK_M

+

C_S⁰

I_Sand I_kbeing deﬁned by eqs 1 and 2, respectively. From eq 5,

D_S(C⁰_S- [S]_{x ) 0}) ) D([P]_{x ) 0}- [P]_{x ) δ}

ﬁnally leading to

[P]_{x )0}) [P]_{x ) δ}

K_M

)

+

2

I_k

I_S

K_M

C_S⁰

I_k

I_S

I_k

I_S

1 +

+

-

1 +

+

- 4

D_SC⁰_S

C_S⁰

(

)

ꢀ

(6)

D

2

As before, an interesting limiting situation, reached upon

increasing the bulk substrate concentration, is when the process

is solely governed by a Michaelis–Menten saturated kinetics

with no interference of substrate diffusion. Then

k₂Γ⁰_Eδ

[P]_{x ) 0}) [P]_{x ) δ}

+

(7)

D

Scheme 4

With conventional cells, the volume-to-electrode surface area

ratio is large, that is, (see Scheme 4) (DRT/FV)^1/2, δ , V/S.

If the last condition is such that (V/S)(δ/D) is large as compared

to the time during which the experiment is carried out, the

building up of P in the solution is negligible (see section 1.3):

[P]_x)δ) 0

and

[P]_{x ) 0}

)

2

K_M

C_S⁰

I_k

I_S

K_M

C_S⁰

I_k

I_S

I_k

I_S

1 +

+

-

1 +

+

- 4

D_SC⁰_S

D

(

)

ꢀ

2

(50) Savéant, J.-M. Elements of Molecular and Biomolecular Electrochem-

istry; Wiley Intersciences: Hoboken, NJ, 2006.

9

J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7263

A R T I C L E S

Limoges et al.

which becomes: [P]_{x ) 0}) k₂Γ_E⁰δ/D under saturated Michaelis–

k₂Γ⁰_Eδ

S

V

[P]_{x ) 0}

)

+

k₂Γ⁰_Et

Menten kinetic control.

D

Under the conditions already deﬁned, the cyclic voltammetric

peak current is simply proportional to [P]_{x ) 0}in all cases:

Under the conditions already deﬁned, i.e., (DRT/FV)^1/2, δ,

the cyclic voltammetric peak current is simply proportional to

[P]_{x ) 0}in all cases:

FV

√

I_p) A × F D

[P]_{x ) 0}

ꢀ_RT

FV

√

I_p) A × F D

[P]_{x) 0}

The proportionality factor A depends on the electron stoi-

chiometry and of the degree of reversibility of the P wave. These

waves, corresponding to a two-electron-two-proton stoichi-

ometry in most cases, are usually not very reversible (for a two-

electron irreversible process with a transfer coefﬁcient of 0.5,

A ≈ 0.07). This sluggish reversibility is especially observed

when proteins are adsorbed on the electrode. The best way for

determining A is to carry out a cyclic voltammetric experiment

with a known concentration of P in the solution, with the same

coating of the electrode surface, but in the absence of the

substrate, that is, in the absence of catalytic reaction.

ꢀ_RT

In summary, introducing the characteristic time (V/S)(δ/D),

the peak current is given by eq 10 in the general case, and by

eq 11 under saturated Michaelis–Menten kinetic control.

D_SC⁰_S

FV

t

√

I_p) A × F D

1 +

×

(

)

+

ꢀ_RT

D

Vδ ⁄ SD

2

K_M

C_S⁰

I_k

I_S

K_M

C_S⁰

I_k

I_S

1 +

+

-

1 +

- 4

(

)

I_S

ꢀ

(10)

(11)

In summary, the peak current is given by:

2

k₂Γ⁰_Eδ

D_SC_S⁰

FV

S

V

I_p) A × F D

+

k Γ⁰t

2 E

FV

√

(

)

√

I_p) A × F D

×

ꢀ_RT

D

ꢀ_RT

D

2

K_M

C_S⁰

I_k

K_M

C_S⁰

I_k

I_S

I_k

I_S

In all cases, the peak current increases linearly with time.

Equation 11 depicts the ampliﬁcation due to accumulation of

the enzyme–product with time and introduces a time constant,

Vδ/SD, which depends on the geometry of the cell deﬁned by

V/S, of the product diffusion coefﬁcient and the thickness of

the steady-state convection/diffusion layer.

1 +

+

-

1 +

+

- 4

(

)

I_S

ꢀ

(8)

2

in the general case and by eq 9 under saturated Michaelis–

Menten kinetic control:

Case b. (DRT/FV)^1/2, V/S , δ. Natural convection no longer

plays a role in P mass transport. The building up of P

concentration is conﬁned within the thickness of the cell. The

peak current is thus given by eq 12 under saturated Michaelis–

Menten kinetic control.

k₂Γ_E⁰δ

FV

√

I_p) A × F D

(9)

ꢀ_RT

D

1.3. Cyclic Voltammetry at Low Volume-to-Surface

Ratios. If the volume (V) to surface (S) ratio is large, as in the

preceding case when conventional cells are used, there is no

point waiting for an increase of the current with time. This is

not the case when the V to S ratio is small. Then P is expected

to accumulate with time leading to an increased sensitivity.

Three cases should be distinguished according to the relative

magnitude of the three characteristic lengths (DRT/FV)^1/2, δ,

and V/S:

k₂Γ⁰_EV

FV

S

V

I_p) A × F D

+

k Γ⁰t

(12)

√

(

)

2

E

ꢀ_RT

DS

In this case too, the peak current increases linearly with time.

S is the electrode surface area and V is the effective volume

within which the diffusion of P actually takes place and may

be different from the total volume of the solution in the cell

(see section 2.4 for an example). It has been implicitly assumed

in the above derivation that an even concentration of P is

established across the whole thickness of the cell at time zero.

In practice this is obviously not the case, leading to a substantial

uncertainty on the location of the time origin, and therefore on

the value of the intercept, in the peak current–time plots

corresponding to eq 12. Analyses based only on the values of

the slope in eq 12 are therefore safer.

Case a. (DRT/FV)^1/2, δ , V/S, as in the preceding case,

but the characteristic time (V/S)(δ/D) (see later) is of the same

order as the experiment duration. The equations governing the

production and mass transport of P are the same as in the

preceding case, except that its bulk concentration builds up

appreciably with time according to

d[P]_{x ) δ}

SD d[P]

) -

(

)

Case c. V/S , (DRT/FV)^1/2, δ. The cell thickness, V/S, is

so small that all the P present in the cell is entirely detected by

cyclic voltammetry, which then gives rise to a surface wave,

which height is proportional to the scan rate rather than to its

square root, according to eq 13 under saturated Michaelis–

Menten kinetic control.

dt

V

dx x ) δ_-

(with reference to Scheme 4, δ_ means approaching δ from

the left).

d[P]_{x ) δ}

[P]_{x ) 0}- [P]_{x ) δ}

SD

V

)

dt

δ

k₂Γ⁰_EV

Equations 6 and 7 are the initial conditions of P building up,

while [P]_x)0- [P]_x)δremains constant over time. Thus

F ²V

S

V

I_p) A′ ×

+

k Γ⁰t

(13)

(

)

2

E

RT DS

SD

Vδ

In this case again, the peak current increases linearly with

time. The ratio between I_pand I_p,0(value of I_pat time t ) 0) is

therefore given by an equation which is formally the same as

[P]_{x ) 0}) ([P]_{x ) 0}- [P]_{x ) δ}) 1 +

t

(

)

where [P]_x)0- [P]_x)δis given by eqs 6 or 7. In the latter case

9

7264 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

Enzyme Bioafﬁnity Electrodes. Direct Detection

A R T I C L E S

electrodes. The latter are particularly valuable in the develop-

ment of electrochemical enzyme immuno- or DNA-assays, their

main advantages being ease of mass production, low-cost, low

background current, good reproducibility, and disposable han-

dling. The screen-printing technology also offers a high ﬂex-

ibility in the design and fabrication of multiplexed arrays of

electrochemical microcells.

At a bare glassy carbon electrode, the cyclic voltammetry of

2,6-DCPAP (scan rate: 0.1 V s^-1) shows a well-deﬁned

diffusion-controlled reversible wave, with a formal potential E°′

(average of the anodic and cathodic peak potentials) of +0.024

V vs SCE, and a peak potential difference of ∆E_p) 68 mV,

pointing to a relatively rapid overall two-electron transfer, which

involves the 2e^-+ 2H⁺oxidation of 2,6-DCPAP into the

corresponding quinonimine, 2,6-DCQI:

Figure 2. Cyclic voltammetry of 0.1 mM 2,6-dichloro-4-aminophenol and

4-aminophenol in Tris buffer (0.1 M, pH ) 9.0) at a bare (blue) or BSA-

coated (red) electrode: (a, b) glassy carbon electrode; (c, d) screen-printed

electrode. Scan rate ) 0.1 V/s.

Under the same conditions, the cyclic voltammetry of PAP

also gives rise to a well-deﬁned two-electron reversible wave,

but with a formal potential slightly more negative than 2,6-

DCPAP (E°′ ) -0.048 V vs SCE) as expected from the

electron-withdrawing effect of the two chlorine and an apparent

heterogeneous electron transfer rate somewhat slower (∆E_p)

88 mV). Coating the electrode with BSA does not change

signiﬁcantly the formal potentials but signiﬁcantly slow down

electron transfer (∆E_p) 115 mV and 276 mV for 2,6-DCPAP

and PAP, respectively) in line with a partial blocking of the

electrode by the adsorbed protein molecules. The results were

qualitatively similar at screen-printed carbon electrodes, except

that electron transfer is further slowed down, which is not

surprising in view of the heterogeneous composition of the

screen-printed carbon electrode surface, made of a random

distribution of closely spaced insulating and conductive micro-

scopic zones. The consequence of this partial blocking of the

electrode surface is an increase of the peak-to-peak separation

in cyclic voltammetry, reﬂecting an apparent slowing down of

electron transfer kinetics.^53,54The better electrochemical char-

acteristics of 2,6-DCPAP as compared to PAP on all types of

electrodes, led us to select 4-amino-2,6-dichlorophenylphosphate

as substrate. This compound was synthesized according to a

two-step procedure. The phenolic group was ﬁrst phosphorylated

with dibenzylphosphite in tetrachloromethane,^55,56and the

resulting dibenzylated phosphate ester was next catalytically

hydrogenated in the presence of palladium on calcium carbonate

(see Experimental Section).

eq 12. The same remarks as in the preceding case concerning

the lack of precision on the value of the intercept and the better

reliability of the slope values apply.

One way of achieving these conditions is to decrease more

and more the cell thickness. A similar situation is expected when

the product P is converted into an electroactive precipitate on

the electrode surface as for example by oxidation of P with Ag⁺

leading to an Ag precipitate, which can be detected and

quantitated electrochemically.⁵¹However, precipitation has to

be quantitative, fast, and harmless for the enzyme functioning

over the whole range of target analyte concentrations, a

condition that seems difﬁcult to achieve with this technique and

other procedures based on precipitation^37,52as will be discussed

on experimental grounds in section 2.4.

1.4. Current, Enzyme Label Surface Concentration, and

Target Analyte Bulk Concentration. In all cases, plotting the

values of Γ⁰_Eversus the bulk concentration of the labeled target

analyte at binding equilibrium provides the recognition isotherm

(for example, but not necessarily, a Langmuir isotherm) and

thus the value of the binding equilibrium constant. Relating

ﬁnally the peak current to the solution concentration of the

labeled target analyte will allow the estimation of the analytical

performances of the method.

2. Illustrating Example: Alkaline Phosphatase Ampliﬁed

Avidin–Biotin Recognition Electrode. 2.1. Selection of the

Substrate/Product Couple. Electrochemical and Enzymatic

Characterization. Stability and Purity of the Reagents. The

electrochemical properties of 2,6-DCPAP were characterized

in cyclic voltammetry both at bare carbon electrodes and at

carbon electrodes coated with an adsorbed layer of bovine serum

albumin (BSA). The resulting voltammograms were also

compared with the curves obtained for PAP under the same

conditions (Figure 2). Two kinds of carbon-based electrodes

were evaluated, glassy carbon and screen-printed carbon

The new substrate was then characterized electrochemically

and enzymatically. Cyclic voltammograms of 2,6-DCPAPP

(Figure 3a) show during the forward scan an irreversible

oxidation peak, and upon scan reversal, the reversible reduction

peak followed by the appearance of a reversible system

(53) Amatore, C.; Savéant, J-M.; Tessier, D. J. Electroanal. Chem. 1983,

147, 39–51.

(54) Osborne, M. D.; Seddon, B. J.; Dryfe, R. A. W.; Lagger, G.; Loyall,

U.; Schäfer, H.; Girault, H. H. J. Electroanal. Chem. 1996, 417, 5–

15.

(55) Silverberg, L. J.; Dillon, J. L.; Vemishetti, P. Tetrahedron Lett. 1996,

37/6, 771–4.

(51) Hwang, S.; Kim, E.; Kwak, J. Anal. Chem. 2005, 77, 579–84.

(52) Patolsky, F.; Katz, E.; Bardea, A.; Willner, I. Langmuir 1999, 15,

3703–6.

(56) Authier, L.; Schöllhorn, B.; Moiroux, J.; Limoges, B. J. Electroanal.

Chem. 2000, 488, 48–58.

9

J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7265

A R T I C L E S

Limoges et al.

4-aminonaphtol and 4-aminonaphtylphosphate (∆E ) 0.46 V).⁵⁹

This is an important advantage for increasing the selective

electrochemical oxidation of 2,6-DCPAP without interference

from the electrochemical oxidation of the 2,6-DCPAPP sub-

strate. The problem of an insufﬁcient discrimination between

the substrate and product detection has already been identiﬁed

with the PAP/PAPP couple as a source of signiﬁcant background

current, hindering the improvement of the detection limit of the

method.³⁵

The ability of alkaline phosphatase to hydrolyze the new

substrate is illustrated in Figure 3b. Incubation of the 1 mM

substrate solution in the presence of an excess of AP (20 nM),

leads to a nearly complete conversion of 2,6-DCPAPP to 2,6-

DCPAP product in less than 5 min. Hydrolysis follows a simple

Michaelis–Menten mechanism. The two pertinent kinetic pa-

rameters k₂and K_Min solution could thus be determined,

spectrophotometrically (k₂) 1837 s^-1, K_M) 48 µM) and

electrochemically (k₂) 1100 s^-1, K_M) 67 µM) by measuring

the substrate dependence of the rate of reaction at pH 9.0 in

0.1 M Tris containing 1 mM MgCl₂, henceforth designated as

TB (see Experimental Section), which correspond to the optimal

conditions for hydrolysis of PAPP by AP.⁶⁰The K_Mis of the

same order as for p-nitrophenylphosphate (K_M) 82 µM),⁶⁰

PAPP (K_M) 56 µM)⁶⁰or phenyl phosphate (K_M) 36 µM).⁶¹

The turnover number is also of the same order as that found

for other aromatic monoester phosphates such as p-nitrophe-

Figure 3. (a) Cyclic voltammograms: [(—) ﬁrst scan, (...) second scan]

recorded at a glassy carbon electrode immersed in a 0.1 M Tris buffer (pH

9.0) containing 1 mM 2,6-DCPAPP and 1 mM MgCl₂. (b) After addition

of 20 nM AP and after a few seconds (blue), 1 min (green), 3 min (red),

and 5 min (magenta). Scan rate ) 0.1 V/s; starting potential ) -0.3 V vs

SCE.

nylphosphate (k₂) 1800 s )

^{-1 62}or phenyl phosphate (k₂) 800

s^-1).⁶¹

The N-AP kinetics was also determined electrochemically at

pH ) 6.5 (MES buffer containing 0.1 wt% BSA and 1 mM

MgCl₂), leading to k₂) 45 s^-1and K_M) 20 µM.

Scheme 5

It has been previously observed that PAP is not very stable

in aerated alkaline solutions.^49,60This instability was ascribed

to a slow decomposition of PAP by air-oxidation, as in several

other hydroquinone and aminophenol derivatives.⁵⁹Despite a

notably higher redox potential and thus a lower susceptibility

to air-oxidation, a similar instability was observed with 2,6-

DCPAP. It leads, in aerated TB, at pH 9.0, to a continuous

decrease of the cyclic voltammetric response with time (Figure

4), which is also related to the presence of dioxygen. The

decrease is not caused by electrode passivation since transferring

the electrode into a fresh solution of 2,6-DCPAP allows

recovering the starting voltammetric response. The rate of

decomposition at neutral pH is considerably slower and almost

independent of the presence of dioxygen. The moderate instabil-

ity of 2,6-DCPAP at alkaline pH is however not critical in the

chronoamperometric technique deﬁned in section 1.2, since 2,6-

DCPAP is generated in the proximity of the electrode surface

and electrochemically oxidized as soon as formed.

(57) It may seem at ﬁrst sight that more electrons are exchanged during

the forward anodic scan than during the reverse cathodic scan. This

is in fact due to the diffusion toward the solution of the anodic scan

product, 2,6-DCQI, during the time elapsed between the anodic and

cathodic peaks, as conﬁrmed by simulation of the voltammogram.

(58) Xu, Y.; Halsall, H. B.; Heineman, W. R. J. Pharm. Biomed. Anal.

1989, 7, 1301–11.

corresponding to the reduction of 2,6-DCQI. These observations

point to the following reaction scheme (Scheme 5).⁵⁷

(59) Másson, M.; Rúnarsson, O.; Jóhannson, F.; Aizawa, M. Talanta 2004,

64, 174–80.

It is worth noting the absence of oxidation wave at ∼0.06 V

during the ﬁrst anodic scan, pointing to a very low level of

residual traces of 2,6-DCPAP in the synthesized substrate. The

difference between the oxidation peak potentials of 2,6-DCPAP

and 2,6-DCPAPP is quite large (∆E ) 0.68 V), larger than

between PAP and PAPP (∆E ) 0.46 V)^49,58or between

(60) Thompson, R. Q.; Barone, G. C., III; Halsall, H. B.; Heineman, W. R.

Anal. Biochem. 1991, 192, 90–5.

(61) Bauer, C. G.; Eremenko, A. V.; Ehrentreich-Förster, E.; Bier, F. F.;

Makower, A.; Halsall, H. B.; Heineman, W. R.; Scheller, F. W. Anal.

Chem. 1996, 68, 2453–8.

(62) Ciana, L. D.; Bernacca, G.; Bordin, F.; Fenu, S.; Garetto, F. J.

Electroanal. Chem. 1995, 382, 129–135.

9

7266 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

Enzyme Bioafﬁnity Electrodes. Direct Detection

A R T I C L E S

Figure 4. (a) Cyclic voltammograms of 2,6-DCPAP (0.1 mM in aerated

0.1 M Tris buffer, pH 9.0) recorded at different time intervals after

immersion of the glassy carbon electrode in the freshly prepared solution.

Scan rate ) 0.1 V/s. (b) Decay of the voltammetric anodic peak current

(from a) as a function of time. Color codes are the same in a and b.

The amount of residual traces of 2,6-DCPAP contained in

the solid substrate powder was estimated in cyclic voltammetry

from the anodic peak current recorded at ∼0.1 V at a glassy

carbon electrode immersed in a series of fresh solutions of 1 or

10 mM 2,6-DCPAPP in TB. Less than 0.05% of 2,6-DCPAP

was estimated to be present in the synthesized 2,6-DCPAPP.

The stability of the substrate 2,6-DCPAPP to spontaneous

hydrolysis in alkaline buffer was also monitored by the evolution

with time of the residual voltammetric peak current correspond-

ing to 2,6-DCPAP in a solution of 1 mM 2,6-DCPAPP in TB.

The fact that it is practically constant after several hours in TB

suggest either that the spontaneous hydrolysis of 2,6-DCPAPP

is very slow or, alternatively, that the slow formation of 2,6-

DCPAP is compensated by its decomposition.

Figure 5. Chronoamperometric current density responses recorded at b-IgG-

coated screen-printed electrodes incubated overnight at room temperature

in solutions containing different concentrations of N-AP. Applied potential:

0.25 V vs SCE. The solution contains a 0.1 M Tris buffer (pH 9.0), 1 mM

MgCl₂and 1 mM of 2,6-DCPAPP. The number on each curve is the N-AP

bulk concentration expressed in picomol/liter.

2.2. Chronoamperometric Responses under Steady State

Natural Convection/Diffusion Conditions. A series of screen-

printed carbon electrodes were coated with a saturated mono-

layer of b-IgG by means of an irreversible adsorption of the

protein on the surface of the electrode. Nonspeciﬁc adsorption

during the N-AP binding step was minimized by a follow-up

blocking of the unoccupied sites on the electrode surface with

bovine serum albumin (BSA). The b-IgG-modiﬁed screen-

printed carbon electrodes, thus prepared, were incubated over-

night in N-AP solutions of different standard concentrations

(C⁰_N-AP) at room temperature. Relatively long times are indeed

required to ensure that the binding of N-AP to the b-IgG-

modiﬁed electrode has reached equilibrium, even at the lowest

N-AP concentrations (it was checked that, for the lowest

concentrations, prolonged incubation times lead to the same

limiting current responses). The binding recognition was also

achieved in a large volume of N-AP solution in such a way

that there is no signiﬁcant N-AP depletion in the analyte solution

(C⁰_N-AP≈ constant). Once equilibrium binding was reached, the

electrodes were washed and immersed in the enzyme–substrate

solution (2 mL of TB containing 1 mM 2,6-DCPAPP) for

steady-state chronoamperometric measurements. Typical chro-

noamperometric responses obtained with a series of b-IgG-

modiﬁed screen-printed carbon electrodes incubated with dif-

ferent standard N-AP concentrations, which cover a range

extending from 2 to 2 × 10⁵picomol/liter are shown in Figure

5. The chronoamperometric measurements were recorded at 0.25

V versus SCE, a potential sufﬁciently positive to efﬁciently

oxidize 2,6-DCPAP but not too high so as to avoid the direct

oxidation of the substrate. The steady-state currents, reached

after some tens of seconds, result from natural convection in

the solution, which leads to the formation of a steady-state diffu-

Figure 6. (a) Variations of the steady-state chronoamperometric current

density responses, at a b-IgG-coated screen-printed electrode, with the

solution concentration of the enzyme-labeled analyte. The currents were

recorded at pH ) 9 (red and blue dots) and pH ) 6.5 (green dots) in

solutions containing 10 mM (red dots) and 1 mM (blue and green dots) of

the substrate 2,6-DCPAPP. Other conditions are the same as in Figure 5.

All data were corrected from the average blank response obtained at

C⁰_N-AP) 0. Error bars represent the standard deviation. (b) Variation of the

surface concentration of the recognized enzyme-labeled analyte as a function

of its solution concentration during incubation. Solid lines are simulations

according to the theoretical relationships of section 1.1 (see text).

sion-convection layer. The thickness of the diffusion-convection

layer may be estimated independently from the steady-state

chronoamperometric response of a standard solution of 2,6-

DCPAP. The steady-state current density being expressed as I

) 2FDC⁰_P/δ, it follows that D/δ ≈ D_S/δ ) (2.1 ( 0.3) × 10^-4

cm/s.

The steady state current densities thus obtained were plotted

against C⁰_N-APin Figure 6 (blue dots), after correction from the

average blank response. The latter was obtained by measuring

the current responses at a series of b-IgG-coated electrodes

immersed in the enzyme–substrate solution without preincuba-

tion in a N-AP solution (C_N⁰_-AP) 0). Repeated experiments

showed that the standard deviation on the determination of the

9

J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7267

A R T I C L E S

Limoges et al.

average blank current density is ∆I ) ( 20 nA/cm². In addition,

it was veriﬁed that the nonspeciﬁc adsorption of N-AP is

rendered negligible by the presence of BSA on the electrode

surface by measuring the current response at electrodes coated

with BSA only and incubated in a N-AP solution of relatively

high concentration (5 × 10^-10M). A nonspeciﬁc current

response corresponding to 0.1% of the speciﬁc current response

was obtained, indicating a relatively low nonspeciﬁc binding.

The resulting curve has roughly the shape of a recognition

isotherm, although the transition between the ascending linear

portion and the horizontal asymptote is steeper than in a

Langmuir isotherm. That it is not a true recognition isotherm

results from another series of similar experiments carried out

at a larger substrate concentration (red dots), 10 mM instead of

1 mM. The ensuing curve does not change upon further increase

of the substrate concentration, showing that the partial control

by substrate diffusion, present in the blue-dot experiments,

vanished in the red-dot experiments as the parameter I_k/I_S)

k₂Γ_E⁰δ/D_SC_S⁰decreases upon increasing C⁰_S.

Figure 7. Michaelis–Menten plot of the steady-state chronoamperometric

current density responses as a function of 2,6-DCPAPP concentrations

recorded at a b-IgG-modiﬁed electrode loaded with a low amount of N-AP

(i.e., obtained after incubation in a solution of 30 pM N-AP). Other

conditions are the same as in Figure 5.

This conclusion falls in line with the observation that the red

and blue dots fall on the same line at lower values of C⁰_N-AP(pM),

hence of Γ⁰_E, which also results in a decrease of the parameter

I_k/I_S. The whole red-dot data and the low-concentration range

blue-dot data may thus be safely treated as representing the

avidin–biotin recognition isotherm on the electrode surface. The

limiting value of the current density at large values of C_N⁰_-AP

corresponds to the maximal enzyme coverage, Γ⁰_E,max. Since the

process is entirely governed by the enzyme kinetics, as just

ascertained, application of eq 2 leads to I ) I_k,max) 2Fk₂Γ_E⁰_,max

and thus to k₂Γ⁰_E,max) 2 × 10^-9mol cm^-2s^-1. Assuming that

biotin with avidin (1.7 × 10¹⁵M^-1) or streptavidin (2.5 × 10¹³

M^-1) in homogeneous solution,^64,65in line with the observation

made with ﬂuorescent-labeled biotins showing that the nature

and length of the structure that links the biotin to the ﬂuorescent

probe affect signiﬁcantly the binding with avidin.^66–68For

example, homogeneous binding constants values ranging from

10⁸to 10¹⁰M^-1have been reported, which is about 5–7 orders

of magnitude lower than the native biotin for avidin.^66–68

A

similar decrease in the afﬁnity for avidin can be anticipated in

the present case for the binding of large biotinylated proteins

such as b-IgG. An additional factor might be a lesser access to

the immobilized biotin ligand at the electrode surface or a steric

hindrance of available sites by bound N-AP.⁶⁹

Further evidence that immobilized N-AP has the same

reactivity as in homogeneous solution was obtained by measur-

ing the Michaelis–Menten constant of the immobilized enzyme.

At low N-AP coverages, as in the experiments reported in Figure

7, eq 4 indeed applies leading to a K_Mvalue of 40 µM. This

value is close to that measured for N-AP in homogeneous

solution (K_M) 48 µM) by means of steady-state spectropho-

tometry (section 2.1).

Coming back to the blue-dot data in Figure 6a, simulation

that takes into account a partial control by substrate diffusion

according to eq 3 leads to a satisfactory ﬁtting of the experi-

mental data using already determined parameters, namely k₂)

1500 s^-1, K_M) 50 µM, D_S/δ ) 2.1 × 10^-4cm/s, K_b) 7 ×

10⁷M^-1, Γ_E⁰_,max) 1500 fmol cm^-2

The applicability of the theoretical relationships may be

further tested with the data obtained at pH ) 6.5 (green dots in

Figure 6) where the enzymatic reaction is expected to be slower

than at pH ) 9. This is indeed the case with the consequence

that, even with a substrate concentration of 1 mM, the enzymatic

kinetics governs the current over the whole range of N-AP

k₂is the same as in homogeneous solution (i.e., k₂) 1500 s^-1

,

taking the average between the spectrophotometric and elec-

trochemical determinations), Γ⁰_E,max) 1500 fmol cm^-2. This

value agrees with the coverage that can be estimated from

protein sizes for a saturated monolayer on a perfectly ﬂat surface

(1700 fmol cm^-2). This result was corroborated by additional

experiments that allow the determination of Γ⁰_E,maxindependently

from k₂. The method is the same as the “depletion method”

formerly used with horseradish peroxidase.⁶³It consists in

depositing a droplet of N-AP solution on the surface of the

b-IgG-modiﬁed electrode and determining spectrophotometri-

cally, after an incubation period, the remaining activity of the

enzyme contained in the droplet. On account of the large ratio

of electrode surface/droplet volume, a sizable amount of N-AP

is depleted from the drop. This amount of enzyme normalized

toward the surface area corresponds to the amount speciﬁcally

bound to the electrode surface. Using this method with a

saturating concentration of N-AP a value of Γ⁰_E,max) 1500 (

300 fmol cm^-2was found in agreement with the previous

determination.

From the value of the proportionality factor, 2Fk₂, between

I_kand Γ⁰_E, and considering that one phosphatase is coupled to

one neutravidin, we may then derive the whole recognition

isotherm as shown in Figure 6b. Fitting with a Langmuir curve

(64) Green, N. Biochem. J. 1963, 89, 585–91.

(65) Bayer, E. A.; Wilcheck, M. Methods Biochem. Anal. 1980, 26, 1–45.

(66) Lo, K. K.-W.; Chan, J. S.-W.; Lui, L.-H.; Chung, C.-K. Organome-

tallics 2004, 23, 3108.

K_bC⁰_N-AP

1 + K_bC⁰_N-AP

Γ⁰_E) Γ⁰_E,max

(14)

(67) Marek, M.; Kaiser, K.; Gruber, H. J. Bioconjugate Chem. 1997, 8,

560–6.

leads to a binding constant, K_b) 7 × 10⁷M^-1. This value is

(68) Loosli, A.; Rusbandi, U. E.; Gradinaru, J.; Bernauer, K.; Schlaepfer,

C. W.; Meyer, M.; Mazurek, S.; Novic, M.; Ward, T. R. Inorg. Chem.

2006, 45, 660–8.

much lower than the constants reported for the binding of native

(63) Limoges, B.; Savéant, J-M.; Yazidi, D. J. Am. Chem. Soc. 2003, 125,

(69) Huang, S.-C.; Stump, M. D.; Weiss, R.; Caldwell, K. D. Anal. Biochem.

1996, 237, 115–22.

9192–203.

9

7268 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

Enzyme Bioafﬁnity Electrodes. Direct Detection

A R T I C L E S

control to diffusion control, whereas it is much smoother in the

red and green curves as expected from curves that derive from

a Langmuir isotherm. This observation brings about a ﬁnal touch

to the excellent agreement between theory and experiment with

the testing example that we have chosen.

2.3. Cyclic Voltammetric Responses at High Volume-to-

Surface Ratios. Cyclic voltammetric experiments carried out at

0.1 V/s in the same conditions as the blue-dot steady state

chronoamperometric experiments reported in Figure 6a are

shown in Figure 8 (in which the previous chronoamperometric

results are also reported for comparison as well as three series

of results obtained with small volume-to-surface cells as

described further on).

As before, the average blank current response (mainly due

to the faradaic current of residual traces of 2,6-DCPAP present

in the substrate solution) was subtracted before plotting the peak

current against the concentration of enzyme-labeled analyte. The

cyclic voltammetry and chronoamperometry curves are closely

parallel, in line with the eqs 3, 3′, and 8. The ratio

Figure 8. (a) Variations of the (blue dots) steady-state or (blue squares,

red dots, squares and triangles) voltammetric peak current density (scan

rate: 0.1 V/s) as a function of C⁰_N-AP. The blue dots and squares were obtained

in a conventional cell (V/S . 10 cm), whereas the red dots, squares, and

triangles were obtained in a small volume-to-surface cell (V/S ) 0.25 cm)

after 1 h of enzyme–product accumulation (red dots) or in a thin-layer

electrochemical cell (V/S ) 0.006 cm) after 20 min (red triangles) and 1 h

(red squares) of enzyme–product accumulation. The enzyme–substrate

solution contains a 0.1 M Tris buffer (pH 9.0), 1 mM MgCl₂, and 1 mM of

2,6-DCPAPP. The oblique straight lines are ﬁtted to the linear part of the

isotherms (with a slope of 1). The horizontal green dotted line represents

3 times the standard deviation of the average blank response. (b) Variation

of the surface concentration of the recognized enzyme-labeled analyte as a

function of its solution concentration during incubation. Black solid line is

a ﬁtting according to a Langmuir isotherm.

I_p

A

FV δ

)

₂ꢀ_RT

I

√

D

is equal to 5.5 for V ) 0.1 V/s, meaning, from the values already

known of all other parameters, that A ≈ 0.6, very close to the

prediction made in section 1.2. The ratio I_P/I could be further

improved by raising the scan rate. However, as already

mentioned in section 2.1, the quality of the response at screen-

printed electrodes rapidly degrades as the scan rate is raised:

the coefﬁcient A decreases and the peak shape eventually

disappears. Glassy carbon electrodes give more satisfactory

results in this connection, as seen in Figure 9, showing that the

peak is proportional to the square root of the scan rate.

2.4. Cyclic Voltammetric Responses at Low Volume-to-

Surface Ratios. Two cell conﬁgurations with a low volume-to-

surface ratio were operated.

(a) In a ﬁrst series of experiments, the production of P from

the enzymatic transformation took place in a droplet of substrate

solution (volume V ) 3 × 10^-2cm³), deposited on top of a

b-IgG-coated screen-printed electrode (surface S ) 0.125 cm²)

that has been previously incubated with a given concentration

of enzyme labeled analyte (see the description of the cell in the

Experimental Section). The voltammogram of the 2,6-DCPAP

produced during the incubation period was then recorded as a

function of time (Figure 10a). As shown in Figure 10a′, whatever

the amount of enzyme label speciﬁcally attached to the working

electrode surface, the peak current increases linearly with time.

This is precisely what is predicted by eqs 10–12. At prolonged

incubation times, it was occasionally observed that the response

tends to be higher than predicted (blue data in Figure 10a′).

This effect is indicative of concentrating effect due to partial

droplet evaporation along the incubation time (during the

enzymatic reaction, humidity was not perfectly controlled

because of repetitive openings of the chamber between each

electrochemical measurement).

From application of eq 12 to the data in Figure 10a′, one

obtains Vδ/SD ) 20 ( 9 minutes, meaning that the current

doubles every 20 min. From this value, with D/δ ) 2 × 10^-4

cm/s (determined in the droplet, according to the same procedure

as before), one obtains V/S ) 0.25 ( 0.1 cm, which is

satisfactorily close to the value calculated from the ratio of the

geometric area of the electrode and the volume of the deposited

droplet (V/S ) 0.24 cm).⁷⁰

Figure 9. Cyclic voltammetric current density responses at a b-IgG-

coated glassy carbon electrode, loaded with a low amount of N-AP (i.e.,

obtained by incubation in a 100 pM solution of N-AP) and recorded as

a function of the scan rate in a standard cell. Other conditions are the

same than in Figure 8.

concentrations with no interference of substrate diffusion unlike

what was observed at pH ) 9. The ﬁtting green curve in Figure

6 was accordingly simply obtained from eq 2, with a value of

k₂) 110 s^-1, 10 times lower than at pH ) 9, close to the value

of the rate constant measured in homogeneous solution at pH

6.5 (k₂) 45 s^-1).

Taking into account this difference in the proportionality

factor between current density and surface concentration of the

enzyme-labeled analyte, all green and red dots and low

concentration blue dots fall on the same Γ⁰_E-C⁰_N-APcurve (Figure

6b), which represents the avidin–biotin recognition isotherm with

at saturation a maximal enzyme coverage of Γ⁰_E,max) 1500 fmol

cm^-2and as binding constant of K_b) 7 × 10⁷M^-1

.

The fact that the horizontal asymptotes of the blue and green

curve in Figure 6 are approximately the same is purely

coincidental. It is interesting to note that, in Figure 6a, the

approach of the asymptote is almost discontinuous in the blue

curve, as expected from the passage from enzyme kinetics

9

J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7269

A R T I C L E S

Limoges et al.

Figure 11. Anodic stripping voltammetry of Ag⁰precipitated on the surface

of different N-AP/b-IgG-modiﬁed electrodes. During the enzyme reaction

the electrode was exposed to a solution containing 1 mM 2,6-DCPAPP

and 2 mM AgNO₃for 30 min: (a) anodic stripping voltammograms for

C⁰_N-AP) 10⁵(blue), 10⁴(red), 10³(green), 10²(magenta, see the zoomed

insert), 0 (black) fM; (b) variations of the peak current density with

C⁰_N-AP. Dotted line is the unity slope expected from the Henry’s law. Scan

rate ) 0.1 V/s.

shows a typical series of voltammetric curves recorded during

the enzyme–product accumulation within the thin-layer elec-

trochemical cell. The anodic peak currents are reported, as

previously, as a function of the enzyme reaction time and for

different concentrations of N-AP during the binding recognition

step (Figure 10b′).

The peak currents obtained after 20 min and 1 h of enzymatic

reaction are reported in Figure 8 leading a new calibration curve

with enhanced sensitivity. Using this strategy, a concentration

as low as 5 fM N-AP was clearly discernible from the blank

response.

As discussed in section 1.3, eq 13 cannot be safely used for

the estimation of V_eff/S in view of the uncertainty on intercept

determination. Instead of the intercept, we therefore resorted to

the voltammetric response, I_p,∞, obtained in a high volume-to-

surface ratio cell for the same analyte concentration, C⁰_N-AP, and

therefore the same value of Γ⁰_E. The value of V_eff/S was then

derived from (see eqs 9 and 13)

Figure 10. Variations of the cyclic voltammetric current density responses

with time after saturation of the electrode with b-IgG, incubation in a

solution of N-AP, and exposition to 1 mM 2,6-DCPAPP in TB: (a, a′)

droplet cell (V/S ) 0.25 cm); (b, b′) thin-layer cell (V/S ) 0.006 cm); (a,

b) cyclic voltammograms for C⁰_N-AP) 3000 and 150 fM, respectively,

recorded at different time intervals (indicated in minutes under the

diagrams) during the enzyme-generated product accumulation; (a′) peak

current density variations for the droplet cell at C⁰_N-AP) 3000 (blue),

600 (red), 150 (green), 0 (black) fM; (b′) peak current density variations

for the thin-layer cell at C⁰_N-AP) 1000 (blue), 150 (red), 20 (green), 5

(magenta), 0 (black) fM. Scan rate ) 0.1 V/s. Straight lines are linear

regressions using eqs 11 and 13.

Several other N-AP standard concentrations were similarly

investigated and the voltammetric peak currents recorded after

one-hour enzyme incubation were reported in Figure 8 as a

function of C⁰_N-AP(red dots), where they are compared with all

the other results. The resulting linear part of the log–log

calibration curve (slope ) 1) is, as expected, shifted toward

lower analyte concentrations, allowing one to reach a lower

detection limit compared to previous approaches.

I_p,∞

V_eff

δ

)

slope_thin-layerD S

where slope_thin-layeris the slope of the diagrams in Figure 10b′.

It follows that (V_eff/S)_thin-layer) 43 ( 20 µm, consistent with

independent estimates leading to a value of ∼60 µm (see

Experimental Section).⁷⁰

(b) In another series of experiments aiming at a further

increase of the analytical performance and at avoiding evapora-

tion of the enzyme–substrate droplet, the enzyme–substrate

solution was conﬁned within a thin-layer between a planar

screen-printed electrochemical cell and a glass coverslip for

microscope slide. This was simply done, just prior to the enzyme

reaction, by spreading uniformly, with the help of the glass

coverslip, a small droplet of substrate solution (4 µL) over all

three electrodes (see the description of the cell in the Experi-

mental Section). The adhesive forces between the liquid, the

planar electrochemical cell, and the glass coverslip then create

an even thickness of liquid of ∼60 µm on top of the working

electrode, meaning that the ratio of the effective volume to the

electrode surface (V_eff/S) is of the order of 60 µm. Figure 10b

(c) As indicated in section 1.3, precipitating Ag⁰from the

reaction of P with Ag⁺, and recording its oxidation cyclic

voltammogram as function of time, should in principle lead to

a further improvement of the detection limit. The results of such

experiments, carried with the same enzyme-labeled recognition

system as in the preceding experiments are summarized in

Figure 11. Two points are worth noting. One is that in the

10²–10⁵fM N-AP concentration range, the current is not

proportional to C_N⁰_-APcontrary to what is expected from the

Henry law relating Γ_E⁰to C_N⁰_-APin the low N-AP concentration

(70) The slow decomposition of 2,6-DCPAP mentioned in section 2.1

(Figure 4) does not affect these results by more than a factor of 20%

in the case of the droplet or 5% in the case of thin-layer cell.

9

7270 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

Enzyme Bioafﬁnity Electrodes. Direct Detection

A R T I C L E S

range, a behavior that is indeed observed in all our experiments

(see Figure 8). The same type abnormal behavior appears in

the experiments, dealing with DNA hybridization, presented for

the introduction of the method,⁵¹thus revealing the existence

of an artifact in the relationship between the production of P

and its translation as an Ag⁰precipitate. As second worrying

observation is that there was no detectable signal for a N-AP

concentrations below 100 fM. These observations, and similar

ones that can be made with other precipitation techniques,^37,52

shed some doubts about their reliability.

3. Discussion of the Analytical Performances. Having tested

the applicability of the theoretical relationships established in

section 1, we may now use them to estimate the analytical

performances of the approaches that we have envisaged, namely,

steady state chronoamperometry; immediate cyclic voltammetry

of the electroactive product generated under steady state

conditions in a conventional cell; and delayed cyclic voltam-

metry of the electroactive product generated under steady state

conditions in a small volume-to-surface area ratio, which allow

its building up with time.

fmol/cm²).⁷²Compared with other electrochemically detectable

enzyme labels, especially redox enzyme labels (Scheme 1a) for

which a relationship equivalent to eq 16 applies,⁷³the N-AP

label leads to a lower detectable enzyme label surface concen-

tration because its turnover number for 2,6-DCPAPP (1500 s^-1

)

is higher than the turnover measured for redox enzyme labels

such as horseradish peroxidase (280 s^-1) in the presence of the

[Os(bpy)₂pyCl]^2+/+mediator,⁶³or as glucose oxidase toward

its glucose substrate (600 s^{-1 74}

or biotinylated diaphorase

)

toward its NADH substrate (700 s^-1).⁷³As just recalled, in the

case of redox enzyme labels the current response is entirely

governed by the enzyme kinetics with no dependence upon the

kinetics of mass transport. This is the reason that the sensitivity

enhancement upon downscaling from a macroelectrode to an

ultramicroelectrode reported in references 75, 76, and 77 seems

puzzling.

As regards the lowest detectable concentration of the target

analyte in solution we ﬁnd, dealing with the avidin–biotin

recognition, by application of eq 15 (with Γ⁰_label,max) 1500 fmol

cm^-2and K_b) 5 × 10⁷M^-1, see section 2.2) a detection limit

of the order of 2 pM. A concentration of 2 pM was indeed

experimentally detected (see Figures 6 and 8).

In all three cases, looking for the smallest detectable

concentration of the analyte, C_N⁰_-AP, and, accordingly, to the

smallest surface concentration of the recognized analyte Γ_E⁰

implies that the enzyme kinetics controls the current response

and that the system lies in the linear portion of the current

3.2. Cyclic Voltammetry in High Volume-to-Surface Ratio

Cells. Using cyclic voltammetry to detect the electroactive

product of the enzymatic reaction, instead of the steady state

chronoamperometry, allows an increase of the sensivity, that

is, the ratio between the current response and the surface

concentration or the incubation concentration of the labeled

analyte (we note incidentally that this enhancement does not

apply to redox enzyme labels involving a redox-mediated

catalytic response). As seen in Figure 8, a typical value for this

enhancement is a factor of 5. The average blank voltammetric

current response can thus be determined with an improved

precision. This is true if the uncertainty on this determination

remains of the same order as in chronoamperometry. This is

indeed the case, since an absolute uncertainty of ∆I ) 25 nA/

cm²was found from the average cyclic voltammetric current

recorded in a 1 mM 2,6-DCPAP (scan rate: 0.1 V/s), which is

very close to the chronoamperometry value. Therefore the

detection limit is expected to reach Γ⁰_label,min) 40 amol cm^-2

and C_a⁰_nalyte,min) 0.4 pM, with the avidin–biotin recognition

system.

density: C⁰_N-APor Γ_E⁰-C_N⁰_-APcurves, where: I or I_p

k₂Γ⁰_E. It

should be borne in mind that comparing the analytical perfor-

mances of several heterogeneous bioafﬁnity assay methods

implies comparing the lowest detectable surface concentration

of the label that has been speciﬁcally immobilized on the solid

phase, Γ⁰_label, and not directly the lowest concentration, C⁰_analyte

,

that can be detected in the incubation solution. Indeed, in the

Henry region of the recognition isotherm (eq 14), the propor-

tionality between Γ⁰_labeland C_a⁰_nalytedepends on the recognition

binding constant and the label maximal coverage, according to:

Γ⁰_label) Γ⁰_label,maxK_bC_a⁰_nalyte

(15)

and so for the proportionality between the current and C_a⁰_nalyte

.

For the same performance in terms of Γ⁰_label, the performance

in terms of C⁰_analytewill be the better the larger the recognition

binding constant is.

3.1. Simple Steady-State Chronoamperometry in Conven-

tional Cells. Under these conditions, and assuming that the

lowest detectable current response corresponds to 3 times the

standard deviation of the blank current response, the detection

limit of the label coverage, Γ⁰_label,min, is then given by:

3.3. Accumulated Cyclic Voltammetry in Small Volume-to-

Surface Ratio Cells. Product accumulation monitored by cyclic

voltammetry, in small volume-to-surface ratio cells, is expected

to result in further improvements. That this is indeed the case

is seen in Figures 8 and 10, the detection limits being sum-

marized in Table 1.

With the thinnest cell arrangement, after 1-hour reaction time,

a value of Γ⁰_label,minas low as 0.2 amol/cm²could be reached,

corresponding to a value of C⁰_analyte,min) 2 fM for the

avidin–biotin recognition system. These values compare favor-

ably with those achieved by the most performing current

techniques.

∆I

2Fk₂

Γ⁰_label,min) 3 ×

(16)

A key-factor of the analytical performance is thus the enzyme

turnover. The other factor, ∆I, is the uncertainty on the

measurement of the blank current density. From repeated

experiments, a reasonable estimate is ∆I ) 20 nA/cm². Since,

k₂) 1500 s^-1, Γ_l⁰_abel,min) 0.2 fmol cm^-2, a performance that

compares favorably with advanced ﬂuorescence techniques (2

ﬂuorophores/µm², i.e. 0.3 fmol/cm²)⁷¹or nanoparticle-ampliﬁed

surface-plasmon resonance (8 × 10⁸molecules/cm², that is, 1.3

For example, with an AP label and the PAPP substrate, a

detection limit of 54 fM of mouse IgG was found²⁷with a

(73) Limoges, B.; Marchal, D.; Mavré, F.; Savéant, J.-M. J. Am. Chem.

Soc. 2006, 128, 2084–92.

(74) Bourdillon, C.; Demaille, C.; Guéris, J.; Moiroux, J.; Savéant, J.-M.

J. Am. Chem. Soc. 1993, 115, 12264–9.

(71) Lehr, H.-P.; Reimann, M.; Brandenburg, A.; Sulz, G.; Klapproth, H.

Anal. Chem. 2003, 75, 2414–20.

(75) Zhang, Y.; Pothukuchy, A.; Shin, W.; Kim, Y.; Heller, A. Anal. Chem.

2004, 76, 4093–7.

(72) He, L.; Musick, M. D.; Nicewarner, S. R.; Salinas, F. G.; Benkovic,

S. J.; Natan, M. J.; Keating, C. D. J. Am. Chem. Soc. 2000, 122,

9071–7.

(76) Zhang, Y.; Kim, H. H.; Heller, A. Anal. Chem. 2003, 75, 3267–9.

(77) Xie, H.; Zhang, C.; Gao, Z. Anal. Chem. 2004, 76, 1611–7.

9

J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7271

A R T I C L E S

Limoges et al.

fM for a 27-base length oligonucleotide.⁸⁴In these two cases,

the Γ⁰_label/C⁰_analyteratio is most probably not very different from

our case,⁸⁵indicating that our approach can potentially reach

similar, if not better, analytical performances in DNA hybridiza-

tion assays.

En route toward a further decrease of the cell thickness aiming

at a further improvement of the detection limit, it is tempting

to embark on a product precipitation strategy, which is indeed

equivalent in principle to a decrease of the cell thickness. How-

ever, as seen in section 2.4, these approaches lack reliability.

Table 1. Detection Limits in Terms of Surface Concentration of the

Enzyme Label and Solution Concentration of the Labeled Analyte for

the Avidin–Biotin Recognition System

Γ⁰_label,min

C⁰_analyte,min

(fM)

method

(amol cm^-2

)

steady-state chronoamperometry in a

standard cell (V/S >> 10 cm)

immediate cyclic voltammetry^ain a

standard cell (V/S >> 10 cm)

cyclic voltammetric^adetection after

accumulation in small volume-to-surface

ratio cells:

200

40

2000

400

Concluding Remarks

droplet (V/S ) 0.25 cm) with 1 h

13

130

5

accumulation

The theory of enzyme-ampliﬁed bioafﬁnity electrodes with

electrochemical product detection has been established for

several modes of detection. In the steady state chronoampero-

metric mode, product detection consists in measuring the steady-

state current after imposition of a potential to the electrode

sufﬁciently positive for the product to be immediately oxidized,

preventing its diffusion toward the solution. The theoretical

relationships that then link the electrochemical current response

to the amount of recognized labeled target analyte on the surface

are governed by two normalized parameters that measure the

Michaelis–Menten competition in the enzyme, on the one hand,

and the competition between the enzymatic kinetics and the

diffusion of the substrate, on the other. The electrochemical

response is eventually related to the labeled target analyte

concentration in solution through the recognition isotherm.

Application of this relationship offers a way to determine its

characteristics, maximum coverage, and recognition binding

constant. Guided by the theoretical relationships, careful analysis

of the experimental data allows elimination of the interference

of substrate diffusion so as to derive a true recognition isotherm.

Further theoretical relationships have been derived for cyclic

voltammetric detection of the electroactive product generated

under steady-state diffusion-natural convection conditions, in

standard cells or by accumulation of the enzyme–product in

cells of large surface/volume ratios. The same two normalized

parameters govern the peak current density. This is ampliﬁed

as compared to the chronoamperometric response by a factor

that increases with the scan rate. In cells of small volume/surface

ratios, the peak current increases linearly with time, which

provides an additional source of ampliﬁcation. Although less

sensitive, the voltammetric detection of the enzyme-generated

product without accumulation has the advantage of being well-

suited to electrochemical detection systems involving array-

based bioafﬁnity electrodes for multianalyte afﬁnity assays.⁵

The avidin–biotin recognition process in a system that

involves alkaline phosphatase as enzyme label and 4-amino-

2,6-dichlorophenyl phosphate as substrate, generating 4-amino-

2,6-dichlorophenol as electrochemically active product has

proved to be particularly pertinent for testing the theory. The

agreement between theoretical predictions and experiment was

found excellent in all detection regimes.

thin-layer (V_eff/S ) 0.0043 cm) with 20

min accumulation

thin-layer (V_eff/S ) 0.0043 cm) with 1 h

accumulation

0.5

0.2

2

^aScan rate: 0.1 V/s.

sandwich-type immunoassay performed in a self-contained

microelectrochemical cell using cyclic voltammetry (50 µm

diameter × 8 µm deep cavity with individually addressable

electrodes, and a V/S ratio of 0.1 cm during the enzyme

reaction). This result is consistent with our 130 fM N-AP

detection limit obtained by cyclic voltammetry in a droplet (V/S

) 0.25 cm), if we assume that the Γ⁰_label/C_a⁰_nalyteratio is of the

same order in both cases. The miniaturization of the electro-

chemical cell achieved in this study allows the detection of lower

absolute amount of analyte. However, it does not provide per

se any improvement of the detection limit. This is rather

governed by the V/S parameter. Its further decrease to a value

of 45 µm thus allows the detection of an analyte concentration

as low as 2 fM. Although less simple to implement, it is worth

noting that similar low V/S value can be reached in a micro-

ﬂuidic channel incorporating electrochemical detection.^78,79

The detection limits achieved in the present work also

compare well with the most sensitive labeling techniques

developed so far, such as for example, silver-enhanced gold

nanoparticle-label which were detected with an image analyzer

at a surface concentration as low as 0.0025 nanoparticle-label/

µm², that is, 0.5 amol/cm².^80–82The femtomolar detection limit

of our simple approach compare also favorably with sophisti-

cated highly sensitive bioafﬁnity assays based on electrogen-

erated chemiluminescent (ECL) detection of DNA hybridization,

with which a 1 fM concentration of a short oligonucleotide target

could be detected.⁸³To reach such a low concentration, the

method takes advantage of polystyrene microbead labels loaded

with a large amount of [Ru(bpy)₃]²⁺(∼7.5 × 10⁹molecules

per bead) which, following the binding event, was released in

an acetonitrile solution for ECL measurement. Another approach

based on the encapsulation of a large amount of a ﬂuorophore

within a silica nanoparticle label led to a detection limit of 0.8

(84) Zhao, X.; Tapec-Dytioco, R.; Tan, W. J. Am. Chem. Soc. 2003, 125,

11474–5.

(78) Rossier, J. S.; Girault, H. H. Lab on a Chip 2001, 1, 153–7.

(79) Ronkainen, N. J.; Thomas, J. H.; Halsall, H. B.; Heineman, W. R.

Trends Anal. Chem. 2002, 21, 213–25.

(85) Afﬁnity binding constants of 10⁷-10⁹M^-1, with a maximal binding

coverage of few pmol/cm², have been found for heterogeneous DNA

hybridization of short (∼20–25 bases) complementary oligonucleotide

sequences on solid surface.^{71,72,86–88}

(80) Taton, T. A.; Mirkin, C. A.; Letsinger, R. L. Science 2000, 289, 1757–

60.

(81) Bao, Y. P.; Wei, T. F.; Lefebvre, P. A.; An, H.; He, L.; Kunkel, G. T.;

Mu¨ller, U. R. Anal. Chem. 2006, 78, 2055–9.

(86) Okahata, Y.; Kawase, M.; Niikura, K.; Ohtake, F.; Furusawa, H.;

Ebara, Y. Anal. Chem. 1998, 70, 1288–96.

(82) Storhoff, J. J.; Marla, S. S.; Bao, P.; Hagenow, S.; Mehta, H.; Lucas,

A.; Garimella, V.; Patno, T.; Buckingham, W.; Cork, W.; Müller, U. R.

Biosens. Bioelectron. 2004, 19, 875–83.

(87) Niemeyer, C. M.; Burger, W.; Hoedemakers, R. M. J. Bioconjugate

Chem. 1998, 9, 168–75.

(88) Fang, S.; Lee, H. J.; Wark, A. W.; Corn, R. M. J. Am. Chem. Soc.

2006, 128, 14044–6.

(83) Miao, W.; Bard, A. J. Anal. Chem. 2004, 76, 5379–86.

9

7272 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

Enzyme Bioafﬁnity Electrodes. Direct Detection

A R T I C L E S

Scheme 6. Synthesis of Zwitterionic Form of 4-Amino-2,6-dichloro-phenyl Phosphate ^a

^a(i) CCl₄, HPO(OBn)₂, NEt₃, CH₃CN; (ii) H₂, 10% Pd/CaCO₃, CH₃OH.

of pH 7.4 was prepared from 4.3 mM NaH₂PO₄, 15.1 mM

Na₂HPO₄, and 50 mM NaCl. A 0.1 M Tris buffer (TB) of pH 9.0

containing MgCl₂was prepared from 5.54 mM Trizma base (Sigma)

and 3 mM Trizma, HCl (Sigma), and 1 mM MgCl₂·6H₂O. All

aqueous solutions were prepared with water puriﬁed by a Milli-Q

water puriﬁcation system from Millipore. Concentrated stock

solutions of 2,6-DCPAP were prepared in absolute ethanol. Diluted

standard solutions of 2,6-DCPAP were prepared in TB and used

immediately.

Synthesis of 4-Amino-2,6-dichlorophenyl Phosphate. The 4-

amino-2,6-dichlorophenylphosphate was prepared in two steps from

the commercially available 2,6-dichloro-4-nitrophenol, as depicted

in Scheme 6. During hydrogenation, both the protecting benzyl

groups were removed and the nitro group was reduced to the free

amine. Taking advantage of the low solubility of the zwitterion, a

pure monoester phosphate was ﬁnally obtained.

(4-Amino-2,6-dichlorophenyl)-dibenzyl Phosphate Ester. The

dibenzyl phosphate ester was prepared according to a slightly

modiﬁed procedure of the general method given by Silverberg.⁵⁵

The starting compound 2,6-DCPAP (1.0 g, 4.81 mmol) was

dissolved in anhydrous acetonitrile (20 mL) under argon atmo-

sphere. At -10 °C, tetrachloromethane (2.32 mL, 24.04 mmol),

N,N-diisopropylamine (1.76 mL, 10.10 mmol) and N,N-dimethyl-

aminopyridine (59 mg, 0.48 mmol) were added successively under

stirring. Then dibenzyl phosphite (1.70 mL, 7.69 mmol) was added

dropwise. After 2 h, the reaction was complete as determined by

thin-layer chromatography. Then, 0.5 M aqueous KH₂PO₄(20 mL)

was added and the reaction mixture was brought to room temper-

ature. The mixture was extracted three times with ethyl acetate.

The combined organic phase was washed successively with water

and brine, dried over magnesium sulfate, and concentrated in vacuo.

The crude product (2.57 g) was puriﬁed by recrystallization from

ethanol (40–50 mL). The pure dibenzyl phosphate ester was

collected as uncolored needles in 71% yield (1.60 g, 3.42 mmol).

¹H NMR (CDCl₃, 250 MHz) δ/ppm 8.12 (s, 2H, aromatic), 7.27

(s, 10H, phenyl), 5.16 (d, 4H, J ) 8.4 Hz, P-CH₂). ³¹P NMR

(CDCl₃, 250 MHz) δ/ppm -7.74. ¹³C NMR (CDCl₃, 250 MHz)

δ/ppm 149.23, 144.43, 135.04, 134.93, 129.56, 128.88 (CH), 128.69

(CH), 128.08 (CH), 124.43 (CH), 70.94 (CH₂), 70.85 (CH₂). Anal.

Calcd for C₂₀H₁₆Cl₂NO₆P: C, 51.30; H, 3.44; N, 2.99. Found: C,

51.26; H, 3.53; N, 2.89.

The enzyme–product, 2,6-DCPAP, has the further advantage

to be reversible even at an electrode partially blocked with

protein layers. Such a well-deﬁned electrochemical reversibility

is particularly attractive for improving the sensitivity by

enzyme^62,89or electrochemical^28,35recycling of the enzyme-

generated redox product as discussed in the following compan-

ion paper.

Taking the various theoretical factors into account, the

analytical performances could ﬁnally be discussed in a system-

atic way so as to rationally and realistically estimate the

detection limits for all regimes. The values thus reached are

summarized in Table 1. The lowest value was achieved with

cyclic voltammetric detection after accumulation in a thin-layer

cell of 40 µm thickness. Then a value of 0.2 amol/cm²is reached

after 1-hour accumulation, corresponding to 2 fM of target

analyte in solution for the avidin–biotin system.

These performances compare favorably not only with those

of other electrochemical techniques but also with those of

sophisticated ampliﬁcation techniques coupled with silver-

enhanced gold nanoparticle-label detected with an image

analyzer^80–82or with ﬂuorescent-based detection of nanoparticle-

label loaded with a large amount of ﬂuorophores.^84,90

Another clear advantage compared with other high-sensitivity

methods is the linearity of calibration curve (slope 1 in log–log

scale) over several orders of magnitude, providing a sound basis

for precise and sensitive quantitation.

Finally, the analytical performances of the method can also

be improved by multiplying the number of enzymes per

conjugate. In the present study, there is only one AP label per

avidin, but it is possible to replace it by large carrier labels such

as carbon nanotubes or microparticles loaded with numerous

enzymes. It has for example been shown that several thousands

of AP can be attached to a carbon nanotube.⁹¹Several orders

of magnitude improvement in sensitivity could thus be obtained

by combining this strategy with our electrochemical approach.

Likewise combination with a second enzyme should also lead

to further improvements as demonstrated in a preliminary

communication.⁸⁹

2,6-Dichloro-4-aminophenyl Phosphate. A solution of (4-

amino-2,6-dichlorophenyl)-dibenzyl phosphate ester (0.96 g, 2.05

mmol) in methanol (140 mL) was ultrasonicated for 10 min and

purged with argon. Ten-percent palladium on calcium carbonate

(1 g) was added and 1 bar of hydrogen pressure was applied for

12 h under vigorous stirring at room temperature. The reaction

mixture was purged with argon and 1% aqueous ammonia (20 mL)

was added to completely dissolve the product. After ﬁltration over

cellite (2 washings with a solution of 1% ammonia in 5 mL water/

10 mL methanol mixture) the ﬁltrate was titrated with 0.5 M sulfuric

acid until pH 4. The mixture was concentrated under vacuum until

total evaporation of methanol and then kept for 1 h at 0 °C. The

precipitate was ﬁltered off, rinsed two times with cold water, and

dried at 50 °C under vacuum for 3 h affording 322 mg colorless

needles of the zwitterionic form of 4-amino-2,6-dichlorophenyl

phosphate (1.36 mmol, 67% yield). ¹H NMR (D₂O, 250 MHz)

Experimental Section

Reagents. p-Aminophenol (Prolabo), 2,6-dichloro-4-aminophe-

nol, p-nitrophenyl phosphate (Sigma), bovine serum albumine

(BSA) (Sigma), chrompure biotinylated rabbit IgG (Pierce), and

neutravidin-conjugated alkaline phosphatase (Pierce, according to

the supplier one phosphatase is coupled to one molecule of

neutravidin) were used as received. The [Os(bpy)₂pyCl]PF₆was

obtained as previously described.⁶³A phosphate buffer saline (PBS)

(89) Limoges, B.; Marchal, D.; Mavré, F.; Savéant, J.-M. J. Am. Chem.

Soc. 2006, 128, 6014–5.

(90) Matsuya, T.; Tashiro, S.; Hoshino, N.; Shibata, N.; Nagasaki, Y.;

Kataoka, K. Anal. Chem. 2003, 75, 6124–32.

(91) Wang, J.; Liu, G.; Jan, M. R. J. Am. Chem. Soc. 2004, 126, 3010–1.

9

J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7273

A R T I C L E S

Limoges et al.

determined in cyclic voltammetry at 20 °C and using a glassy carbon

electrode. A value of 4 × 10^-6cm²s^-1was found. To check the

electron transfer rate through a protein ﬁlm adsorbed on the

electrode, screen-printed carbon electrodes and glassy carbon

electrodes were coated with a monolayer of BSA by dipping the

electrode in a 0.1 wt % BSA solution for 15 min.

Preparation of b-IgG-Modiﬁed Carbon Electrodes and

Afﬁnity Binding of N-AP. Unless otherwise stated, all experiments

were performed at room temperature. Drops of 0.5 mg mL^-1

biotinylated rabbit IgG in PBS were locally deposited on the sensing

area of the working screen-printed carbon electrodes (drops of 1.5

µL in the case of the 0.0176 cm²electrode and 10 µL in the case

of the 0.125 cm²Andcare’s working electrode were used) and

incubated for 2 h in a water-saturated atmosphere. After thorough

washing with PBS, b-IgG-modiﬁed carbon electrodes were dipped

in PBS containing 0.1 wt % BSA for 20 min and next in a solution

of N-AP diluted in TB (0.5 mL for the 0.0176 cm²screen-printed

carbon electrodes and 5 mL for the Andcare electrodes) for at least

12 h (the N-AP concentration was ranged from 5 × 10^-15to 2 ×

10^-7M). For control experiments, the ﬁrst layer of b-IgG was

replaced by BSA. Once prepared, enzyme electrodes were rinsed

and stored in PBS at 4 °C until used.

For experiments involving accumulation of the enzyme-generated

product in a low volume-to-surface cell two conﬁgurations have

been used (Figure 12). Both were based on planar Andcare screen-

printed electrochemical cell. The ﬁrst one consists of a deposit of

a small droplet of 30 µL of enzyme–substrate solution (1 mM 2,6-

DCPAPP in TB) over the screen-printed electrochemical cell (Figure

12) and monitoring the accumulated product periodically, from a

few seconds to 1 h, by cyclic voltammetry. During the enzyme

reaction, the electrochemical cell was in a chamber saturated with

humidity to minimize evaporation of the droplet. The second

conﬁguration consists of conﬁning 4 µL of the enzyme–substrate

solution (1 mM 2,6-DCPAPP in TB) between a planar screen-

printed electrochemical cell and a glass coverslip for microscope

slide (Figure 12) and, as for the droplet conﬁguration, monitoring

the accumulated product by cyclic voltammetry. The thickness of

thin-layer cell was indirectly determined from the cyclic voltam-

metry of a solution of [Os(bpy)₂pyCl]⁺(0.1 mM). Upon rising the

scan rate, the transition in voltammetry from a thin-layer to a semi-

inﬁnite diffusion layer (l ) 4 × 0.446(DRT/FV)^1/2) allows us to

calculate a ﬁlm thickness of 60 µm (using for the [Os(bpy)₂pyCl]⁺

a diffusion coefﬁcient of 4.5 × 10^-6cm²s^-1). This measured value

is in agreement with the ﬁlm thickness of l ) 63 µm that can be

calculated from the geometrical dimensions of the electrochemical

microcell (9-mm diameter) and the 4 µL of liquid deposited within

the cell.

For experiments involving the reductive precipitation of silver

induced by the enzyme-generated product, the N-AP/b-IgG-

modiﬁed screen-printed carbon electrodes (1.5-mm diameter),

prepared as above with different N-AP coverage, were immersed

in a Tris buffer of pH 9 (adjusted with H₂SO₄) containing 2 mM

MgSO₄, 1 mM PAPP, and 2 mM AgNO₃for 30 min in a black

room.⁵¹After a thorough washing, the N-AP modiﬁed electrodes

were scanned by anodic stripping voltammetry in a 2 N H₂SO₄.

Electrochemical Characterization of 2,6-DCPAPP. A mass

(1.27 mg) of the 2,6-DCPAPP powder was dissolved in water (57.72

µL). A 50 µL portion of this solution was introduced in the

electrochemical cell containing 2.45 mL of TB and 5 µL of a 10.5

µM N-AP solution. After 5 min of enzyme reaction, a cyclic

voltammogram was recorded showing the typical reversible wave

of 2,6-DCPAP. Considering that all of the 2,6-DCPAPP was

hydrolyzed into 2,6-DCPAP within 5 min, it was possible to deduce

from the magnitude of the reversible peak currents and the diffusion

coefﬁcient of 2,6-DCPAP the initial amount of 2,6-DCPAPP added

into the cell. Taking in account dilution and mass of powder

dissolved, we thus calculated an apparent molecular weight of 329.5

g mol^-1for 2,6-DCPAPP.

Figure 12. Droplet and thin-layer small volume-to-surface ratio cells.

δ/ppm 6.79 (s). ¹H NMR (CD₃OD, 250 MHz) δ/ppm 6.64 (s). ¹H

NMR (pyridin-d5, 250 MHz) δ/ppm 6.89 (s). ³¹P NMR (D₂O, 250

MHz) δ/ppm 0.42 (s). ³¹P NMR (CD₃OD, 250 MHz) δ/ppm -3.84

(s). ³¹P NMR (pyridine-d5, 250 MHz) δ/ppm -2.50 (s). MS-ES

(TOF⁺), m/z (relative intensity, fragment): 514.9 (50%, M+M-1,

dimer), 255.9 (100%, M-1 with ³⁵Cl,³⁵Cl), 257.9 (55%, M-1 with

³⁵Cl,³⁷Cl), 259.9 (7%, M-1 with ³⁷Cl,³⁷Cl), 222 (70%, [M - Cl]).

Anal. Calcd for C₆H₆Cl₂NO₄P: C, 27.93; H, 2.34; N, 5.43; P, 12.01.

Found: C, 29.59; H, 2.77; N, 5.62; P, 11.70.

Electrochemical Experiments. An AUTOLAB potentiostat

(Ecochemie) interfaced to a PC computer (GPES software) was

used for cyclic voltammetry and chronoamperometry. For the

experiments performed in a large volume of substrate solution (V/S

>> 10), small screen-printed carbon electrodes (1.5-mm diameter)

were used as disposable working electrodes. They were prepared

from a commercial carbon-based ink (Electrodag PF-407A, Acheson

Colloid) printed on a sheet of transparency ﬁlm (3M). A manual

screen-printer (Circuit Imprimé Français, Bagneux, France) equipped

with a screen-stencil of 77 threads cm^-1was used. In a few

experiments a freshly polished glassy carbon electrode (3-mm

diameter) was used as working electrode. A classical water-jacketed

electrochemical cell (maintained at 20 ( 0.5 °C) equipped with a

saturated calomel reference electrode (SCE) and a platinum counter

electrode was used. For the experiments carried out with a low

V/S cell conﬁguration, electrochemical microcells (Andcare) inte-

grating three screen-printed electrodes on a planar sheet (i.e., two

carbon electrodes as the working and counter electrodes, and a Ag/

AgCl pseudoreference) were used (Figure 12). In this conﬁguration

the working electrode area was 0.125 cm²and the overall electrodes

were delimited by a circular thick insulating layer (9-mm diameter)

allowing the deﬁnition of an electrochemical microcell of a few

tens of microliters. It is worth noting that in the case of standard

ELISA performed in classical wells of a microtiter plate, the V/S

is of ∼0.2 cm, which is not far from the format of our droplet

deposited in the electrochemical microcell.

The voltammetric curves presented in this work were systemati-

cally referred to SCE. The diffusion coefﬁcient of 2,6-DCPAP was

9

7274 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

Enzyme Bioafﬁnity Electrodes. Direct Detection

A R T I C L E S

Homogeneous Kinetics of N-AP with 2,6-DCPAPP as Sub-

strate. The homogeneous kinetics of the reaction of N-AP with

2,6-DCPAPP were determined both spectrophotometrically and

electrochemically. For spectrophotometric measurements, a UV–

visible Hewlett-Packard 8452 diode array spectrophotometer in-

terfaced with a PC computer was used and the temperature of the

standard quartz cell (1-cm path length) was maintained at 20 °C.

Steady-state kinetics were studied by following at 324 nm the initial

rate of 2,6-DCPAP production (ꢀ_324nm) 3525 M^-1cm^-1) as a

function of the substrate concentration (ranging from 10 to 1500

µM). The reactions were initiated by addition of the 2,6-DCPAPP

substrate to a solution of N-AP (5 × 10^-11M) in TB (pH 9.0)

containing 0.1 wt % BSA. The reciprocal plot of the dependence

of the initial rate on the initial concentration of 2,6-DCPAPP was

ﬁtted to a Lineweaver–Burke plot, and values of k₂) 1837 ( 73

s^-1and K_M) 48 ( 3 µM were determined.

For the electrochemical characterization of the homogeneous

kinetics of N-AP, amperometry at controlled potential was used.

The amperometric responses were recorded at a glassy carbon

electrode (polarized at 0.25 V vs SCE) immersed in a solution

containing 0.05 nM N-AP in TB, 0.1 wt% BSA, and different

concentrations of 2,6-DCPAPP (ranging from 30 to 500 µM) that

were injected immediately after the current was stabilized. The

initial linear increases of the current responses as a function of the

substrate concentration were used to calculate the initial rates of

2,6-DCPAP produced in the bulk of the solution. From the ﬁtting

of the data to a Lineweaver–Burke plot, values of k₂) 1073 (

163 s^-1and K_M) 67 ( 31 µM were obtained. The N-AP kinetics

was also determined amperometrically at a lower pH of 6.5 (MES

buffer containing 0.1 wt % BSA and 1 mM MgCl₂). Under these

conditions, values of k₂) 45 ( 2 s^-1and K_M) 20 ( 4 µM were

found.

Determination of the Enzyme Coverage from Depletion of

N-AP Contained in a Droplet Deposited on the b-IgG-Coated

Electrode. The enzyme coverage was independently determined

using the droplet depletion method previously developed by us for

horseradish peroxidase and extended here for AP.⁶³The following

procedure was applied. Drops of 1 µL containing 0.5 mg mL^-1

b-IgG in TB were deposited on the active surface of a series of

screen-printed carbon electrodes and incubated for 2 h in a water-

saturated atmosphere at 4 °C. The electrodes were then rinsed with

TB and dipped for 15 min in 2 mL TB containing 0.1 wt % BSA.

After another washing with TB, drops of 1.5 µL of 0.05 µM N-AP

in TB containing 0.1 wt % BSA were deposited on the surface of

modiﬁed screen-printed carbon electrodes and then further incubated

for 2 h in a water-saturated atmosphere at 4 °C. Next, each extremity

of the electrodes with their droplet over the surface were immersed

and rinsed for 5–10 min in a vial ﬁlled with 2 mL TB containing

0.1 wt % BSA. The activity of N-AP contained into the vial was

then measured spectrophotometrically by monitoring the increase

of absorbance at 396 nm after the addition of 20 µL of 50 mM

p-nitrophenyl phosphate substrate in TB. The unknown diluted

N-AP concentration was ﬁnally determined from a standard

calibration plot of N-AP ranging from 12.5 to 50 pM. From the

determined concentration, the amount of N-AP immobilized on the

electrode was indirectly inferred. An average N-AP coverage value

of 1.5 ( 0.3 pmol cm^-2, corrected from nonspeciﬁc adsorption,

was ﬁnally estimated. This result is relevant to the theoretical value

that can be roughly estimated from the protein sizes of a packed

monolayer of N-AP (i.e., 1.7 pmol cm^-2for a saturated monolayer).

JA7102845

9

J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7275

Article Doi

Theory and practice of enzyme bioaffinity electrodes. Direct electrochemical product detection

DOI: 10.1021/ja7102845

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja7102845

Full text of DOI:10.1021/ja7102845

Products guided by the article

R&D Labs maybe for 124219-30-3

Relevant to this article

Hot Product