Induced conformational change on ferrocenyl-terminated alkyls and their application as transducers for label-free immunosensing of Alzheimer's disease biomarker

Abstract

Alzheimer's disease is the second most common neurodegenerative illness affecting elderly people. Early diagnosis could help to improve a patient's quality of life. The classical sandwich ELISA-based methods are usually costly and time consuming. In this study, we report the design of a label-free immunosensing platform for the sensitive detection of the ApoE protein as a biomarker for Alzheimer's disease. The immunorecognition event induces conformational changes in ferrocenyl alkyl groups tethered to the superficial gold nanoparticles in the vicinity of the antibody. The heavy antigen/antibody complex enhances the electron-transfer (ET) rate constants by bending the ferrocenyl alkyl chain, bringing the ferrocene closer to the gold surface. Determination of the ET rate constants and the analytical performance studies for a series of devices using ferrocenyl alkyl groups with different chain lengths support the proposed mechanism. The best performances and the highest rate constants are observed with sensors having the most flexible chains. The devices are endowed with a large dynamic range (i.e. 0.13 to >1880 ng mL⁻¹) and excellent selectivity and specificity.

---

1. Introduction

Inherently endowed with several interesting features such as selectivity and sensitivity of analysis, high repeatability, and ease of signal quantification,¹ electrochemical biosensors are in direct line to produce devices for point-of-care applications since they comply with the basic requirements for such applications.² Ferrocene derivatives are good candidates to electrochemically transduce biological events and to regenerate redox-active sites in enzymes by acting as electron mediators because of their low oxidation potentials, fast ET rates and stability in an aqueous medium.³ Moreover, ferrocene has been extensively used to study the ET in model systems such as thiol-terminated self-assembled monolayers (SAMs) of alkyl groups or peptides with an end-capping ferrocene group.⁴ Moreover, ferrocene has been covalently linked via an alkyl spacer to boron-doped diamond electrodes.⁵ Two major mechanisms have been recognized: (i) for dense self-assembled layers, electrons tend to travel through the carbon–carbon bond from the ferrocene group to reach the electrode surface (tunnelling mechanism)⁶ and (ii) electrons migrate by hopping from one site to another, especially in proteins.⁷ Recently, it has been reported that the ET can be enhanced by conformational disorder or can be gated by conformational changes.⁵ Changes in the DNA conformation were also exploited to develop a highly sensitive electrochemical hybridization-based DNA sensor, through the monitoring of ET from ferrocene in a favourable position allowing facile ET and another one impeding it.⁸

There are numerous reports on the use of ferrocene in the immunosensing of cancer biomarkers. For instance, ferrocene and thionine were covalently conjugated to anti-α-fetoprotein (AFP) and α-carcinoembryonic antigen (CEA) antibodies, respectively, and co-immobilized on a glassy carbon electrode modified with gold nanoparticles and conjugated with horseradish peroxidase (HRP) to develop a sensitive amperometric immunosensor for simultaneous determination of AFP antigen and CEA.⁹ The two antigens can be detected in a single run with the same dynamic range (i.e. 0.01–50 ng mL⁻¹) and the same limit of detection (∼0.01 ng mL⁻¹). CEA was also detected in saliva and serum using an electrochemical ELISA-like immunoassay. CEA has been sandwiched between monoclonal α-CEA antibodies, covalently immobilized on polyethyleneimine-wrapped CNTs cast onto screen-printed electrodes, and α-CEA tagged ferrocene carboxylic acid encapsulated liposomes. The addition of CEA induces an increase in the differential pulse voltammetry (DPV) current proportional to the antigen amounts.¹⁰ In another report, ferrocene functionalized-peptide nanowires having a diameter of ca. 100 nm were synthesized, coated with poly(diallyldimethyl-ammonium chloride), conjugated with antibodies and decorated with gold nanoparticles, then used as a detection probe for an ultrasensitive electrochemical sandwich immunosensor for human IgG. The capture probe was immobilized on a gold nanoparticles/graphene composite.¹¹

On the other hand, the apolipoprotein E (ApoE) class of proteins is essential to the metabolism of triglyceride-rich lipoproteins and to the central nervous system.¹² The presence of the ApoE4 isoform has been used as a biomarker for Parkinson's or Alzheimer's diseases.¹³ ApoE levels in the cerebrospinal fluid (CSF) of a patient with Alzheimer's disease (AD) are 3-fold higher than in control subjects (4.66 ± 1.96 μg mL⁻¹ vs. 1.92 ± 1.46 μg mL⁻¹).¹⁴ For blood plasma, these levels are 61.50 ± 22.04 μg mL⁻¹ and 52.31 ± 16.61 μg mL⁻¹, respectively, for AD patients and healthy persons with the APOE ε4/ε3 genotype.¹⁵ In general, ApoE is determined using various immunoassay techniques such as enzyme-linked immunosorbent assay (ELISA), surface plasmon resonance and other sandwich-like immunoassays.¹⁶ These methods still suffer from several disadvantages such as lengthy analysis times, necessity of a second antibody to immunosandwich the antigen, the need for a second antibody labelling and several washing steps. The aforementioned disadvantages can be overcome by a direct immunoassay.

In this study, we report the direct determination of ApoE using an electrochemical immunosensing platform based on the conformational changes occurring on alkyl chains in ferrocenyl alkyl-modified gold nanoparticles as a transducing system during the recognition event. The effects of chain length on the analytical performances and on the ET rate constants have been evaluated using a series of ferrocene derivatives with different chain lengths and ApoE as a model analyte.

2. Experimental

2.1. Materials and reagents

All the chemicals and biochemicals, 1-ferrocenyl methanol (97%), ferrocene carboxylic acid (97%), 2-mercaptoethanol (98%), 4-mercaptobutanol (95%), 8-mercaptooctanol (98%), α-lipoic acid (≥98%), dimethylaminopyridine (DMAP) (99%), dicyclohexylcarbimide (DCC) (99%), glutaraldehyde (GL, 50%), HAuCl₄·3H₂O (99.9%), sodium borohydride (≥98%), human IgG (hIgG, ≥95% from human serum) antibody, ovalbumin (OVA, ≥98%), bovine serum albumin (BSA, 98%), phosphate buffer saline (PBS) in tablet form (PBS stands for 0.1 M, pH = 7.4 PBS solutions, unless otherwise stated), silica gel 60 mesh and fluorescent silica-coated aluminium TLC plates were purchased from Sigma-Aldrich (Germany). Chemicals and biochemicals were used without further purification. Capture monoclonal antibody anti-ApoE (α-ApoE produced in vitro cultures) and purified standard ApoE solutions were purchased from Mabtech AB (Nacka Strand, Sweden).

All voltammetric experiments were performed in PBS using a PC-controlled Metrohm Autolab PGSTAT M204 electrochemical workstation with Nova software (v 1.10) for data collection. Screen-printed carbon electrodes (SPCEs) printed on polyethylene terephthalate (PET) sheets with a semi-automatic screen-printing machine (DEK-248, DEK International, Switzerland) were used to perform the electrochemical experiments. The SPCEs comprise a carbon disk serving as a working electrode (3 mm diameter), a printed Ag/AgCl reference electrode and a carbon wire as a counter-electrode.

Scanning electron microscopy (SEM) images were obtained using field emission SEM (Merlin, CarlZeiss). UV-Visible absorption measurements were carried out on a SpectroQuest 2800 spectrophotometer (UNICO, Spain). A thermostatic centrifuge ScanSpeed 1730R (LoboGene A/S, Denmark) was used to purify the AuNP/antibody bioconjugates. All the solutions were prepared in deionized water (>18.2 MΩ cm⁻¹) produced using the Milli-Q system. NMR spectra were obtained in CDCl₃ using Bruker Advance 300 apparatus at 300 MHz. Chemical shifts are given in ppm with tetramethylsilane (TMS) used as an internal reference.

2.2. Preparation of ferrocene derivatives

Ferrocenyl lipoic acid ester (FcL) was prepared according to literature.¹⁷ The FC_nS derivatives (n = 2, 4, 8) were obtained using Steglich conditions as depicted in Fig. S1 in the ESI.† Ferrocene carboxylic acid (115.02 mg, 0.50 mmol) and 1.2 equivalents of n-mercapto-1-alcohol (0.60 mmol) were mixed under stirring in dichloromethane at 0 °C for 15 min, then 1.2 equivalents of DCC (121.8 mg, 0.60 mmol) and 30% molar of DMAP (18.3 mg, 0.06 mmol) were added. The reaction progress was monitored by TLC. The solution was stirred for 24 hours at room temperature to give 1-ferrocenyl-n-mercaptoalkyl acid esters (FC_nS) in 65–76% yields. The products were purified by silica column chromatography using ethyl acetate and hexane (3 : 7). NMR chemical shifts of the products are given in ESI.†

2.3. Immunosensor fabrication

The cystamine-capped gold nanoparticle solution was prepared according to a literature procedure.¹⁸ Briefly, 1 mL of a freshly prepared cystamine solution (213.0 mM) was added to 100 mL of tetrachloroauric(III) acid solution (1.42 mM). The mixture turned yellow upon addition of cystamine and was then stirred vigorously for 10 min. Furthermore, 40 μL of a freshly prepared sodium borohydride solution (10 mM) was quickly injected into the reaction mixture. After vigorous stirring for 20 min, the solution colour turned wine-red. The resulting cystamine-modified gold nanoparticles (cysAuNPs) were stored in the dark at 4 °C.

SPCE was used as a substrate and was cleaned beforehand by 25 cyclic voltammetry (CV) runs between −1.5 and +1.5 V in a PBS solution, then thoroughly washed with deionized water. The cysAuNPs-modified electrode was then immediately prepared by drop-casting 5 μL of the cysAuNPs suspension on the working electrode and allowing it to dry at room temperature (RT). The surface morphologies of the bare and cysAuNPs-modified SPCEs were investigated by SEM, which showed that the AuNPs were physically adsorbed on the porous carbon surface, as detailed in the Results and Discussion section. 10 μL of 2.5 mM GL was dropped onto the AuNPs modified working electrode area, incubated for 1 hour at 4 °C and washed several times with a PBS solution. The modified electrode was immersed for 3 hours (at 4 °C) in a 100 μg mL⁻¹ solution of monoclonal α-ApoE antibodies, which were covalently immobilized on the surface. The electrode was then thoroughly washed twice with PBS. In the next step, ferrocene was electrochemically immobilized by dropping 25 μL of 10 mM ferrocene derivative solution on the cysAuNP-modified SPCE and a potential of −0.4 V was applied for 360 seconds (Fig. S2†). To avoid any type of non-specific interactions, the modified electrodes were further immersed in 3% BSA solution for 1 h at RT. Finally, the as-prepared immunosensor was washed several times with PBS solution and stored at 4 °C for more than 15 days.

The same experimental procedure was used to immobilize α-ApoE antibodies used for the study of the effect of ferrocene derivative chain lengths on the immunosensing response.

Cyclic voltammetry was used to characterize the first modification step of SPCE with AuNPs by scanning the potential between −0.9 and +1.3 V in 0.5 M H₂SO₄ solution at a scan rate of 0.1 V s⁻¹. The stepwise surface modification was monitored by electrochemical impedance spectroscopy (EIS) at an applied potential of 0.2 V. For this, a 50 μL aliquot of a (0.1 M KCl + 5 mM [Fe(CN)₆]^4−/3−) solution was cast on the electrode surface and then the frequency was swept from 100 MHz to 0.1 Hz. All the EIS data are presented as Nyquist plots (i.e. the reciprocal of the real component of the impedance plotted against the imaginary one).

2.4. Immunoassay and electrochemical detection

α-ApoE/FcL/cysAuNPs/SPCE was incubated for 30 min at RT with 50 μL of different concentrations of the ApoE solutions, followed by washing with PBS before the electrochemical measurements. For all the immunosensing experiments, differential pulse voltammetry (DPV) curves were recorded at a scan rate of 50 mV s⁻¹. The immunosensing event is correlated with the variation in the current response before and after immunoreactions. Therefore, the changes in the current response (Δi) of the immunosensors were evaluated as follows: Δi = i_f − i_i, where i_i and i_f represent the current response of the immunosensors prior to and after the electrode incubation in a given concentration of the antigen. The same experimental procedure was followed for the α-ApoE/FC_nS/cysAuNPs/SPCE incubation with ApoE, performed to study the effect of the ferrocene derivative chain lengths on the immunosensing response and on the ET rate constants.

Chronoamperometry was used to monitor the ApoE immunosensing. Chronoamperograms were recorded at potentials of 0.31, 0.54, 0.57 and 0.61 V, which correspond to the oxidation potentials of the ferrocene derivatives FcL, FC₂S, FC₄S and FC₈S.

3. Results and discussion

3.1. Immunosensor design

The ApoE immunosensor was fabricated in a straightforward manner through the consecutive modification of the substrate, as schematically outlined in Fig. 1. cysAuNPs were first used to coat the working surface of the SPCE. The deposition of cysAuNPs onto the bare electrode was confirmed by SEM. A typical rough surface was observed in the case of the bare electrode (Fig. 2A), while for the cysAuNPs/SPCE, AuNPs were uniformly dispersed, but a few aggregated particles were observed on the carbon surface (Fig. 2B). The average diameter of these nanoparticles was ∼20 nm.

Fig. 1 Schematic representation of the immunosensor preparation: (A) adsorption of the cysAuNPs onto the SCPE surface, (B) biofunctionalization with antibodies reticulated by glutaraldehyde and (C) modification by the ferrocene derivative.

Fig. 2 (A and B) SEM images of the bare SPCE and cysAuNPs-modified SPCE, respectively; CVs (C) and EIS plots (D) of SPCEs performed in 0.1 M KCl containing 5 mM [Fe(CN)₆]^4−/3− after each step of the modification (inset (c) CVs of bare SPCE (black) and cysAuNPs-modified SPCE (red) performed in a 0.5 M H₂SO₄ solution at 100 mV s⁻¹ sweep rate). (1) SPCE, (2) cysAuNPs/SPCE, (3) GL/cysAuNPs/SPCE, (4) FcL/GL/cysAuNPs/SPCE, (5) α-ApoE/FcL/GL/cysAuNPs/SPCE, (6) BSA/α-ApoE/FcL/GL/cysAuNPs/SPCE.

CV and EIS were used to characterize the electrode surface modification. The different CVs are displayed in Fig. 2C. The inset shows the CV recorded between 0 and 1.5 V for bare and cysAuNP-modified SPCEs in a 0.5 M H₂SO₄ solution. No faradic current was observed for the bare electrode, whereas a typical curve related to the presence of a gold layer was obtained for the modified electrode. Three redox peaks can be observed; two during the forward scan (at ca. 0.8 and ca. 1.2 V) and one at 0.6 V during the reverse scan corresponding to the oxidation of gold to gold(III) species and their subsequent reduction to gold(0).¹⁹

The potential was scanned between −0.9 and +1.3 V (Fig. 2C) using [Fe(CN)₆]^4−/3− to probe the surface modification. From the figure, we can conclude that a little gain in reversibility and a better ET are obtained after the surface modification with AuNPs, which are partially lost in the subsequent steps because of the introduction of an electronic transfer-impeding layer formed by glutaraldehyde and the antibodies.²⁰ Furthermore, EIS was used to monitor the immunosensor preparation. Fig. 2D exhibits the Nyquist plots of a series of modified electrodes. It is clearly shown that a relatively large charge-transfer resistance (R_ct) was observed at the bare SPCE (curve 1), which drastically decreased after the nanostructuration of the surface by cysAuNPs (curve 2), in agreement with previous reports.²¹ The immobilization of glutaraldehyde onto cysAuNPs/SPCE was accompanied by a gentle R_ct increase (curve 3). This slight increase was perhaps due to the blocking effect of glutaraldehyde.²² Another prominent increase in the resistance was observed after the introduction of the ferrocene derivative (curve 4). Finally, immobilization of the α-ApoE antibody and BSA induced a further increase in the charge-transfer resistance, as shown in curves (5) and (6), respectively.

The α-ApoE/FcL/cysAuNPs/SPCE immunosensor displayed a reversible CV signal related to the oxidation of the ferrocene moiety to the ferrocenium cation, as displayed in Fig. 3A. Moreover, plotting the oxidation and reduction peak currents vs. the sweep rates showed a linear relationship, suggesting a surface-confined ET process and confirming the immobilization of the ferrocene derivative on the electrode surface (Fig. 3B).²³ 3% BSA solution was used to cover the gold surface to prevent any type of non-specific protein interactions.

Fig. 3 (A) CV of the α-ApoE/FcL/cysAuNPs/SPCE immunosensor recorded in PBS at 500 mV s⁻¹; (B) plot of the current versus the scan rate (i vs. v) (inset: CVs of the biosensor recorded in PBS at various scan rates from 50 to 500 mV s⁻¹); (C) DPV signal increase evidencing ApoE/anti-ApoE recognition in the biosensor after incubation in different ApoE solutions with gradually increasing concentrations (inset: variation of chronoamperometry current after incubation in the same solutions); (D) plot of current increase (Δi) vs. the ApoE amounts and fitting the curve.

3.2. Immunosensing of ApoE biomarker

Immunoassays were performed by incubating the as-prepared immunosensors in a series of ApoE solutions prepared at different concentrations; the change in ferrocene oxidation current before and after the incubation was used to monitor the immunoreaction. Immunosensor response to ApoE was calibrated using DPV rather than CV, as the former was found to be more sensitive and reproducible. DPV voltammograms (Fig. 3C) exhibited a gradual increase in Δi values with increasing concentration of ApoE protein in a linear fashion (cf. Fig. 3D). We speculate that the current variation (Δi) is induced, somehow, by the antigen–antibody immunocomplex formation (vide infra for a proposal of the mechanism). For instance, a marked increase in Δi was found to be ca. 4.87 μA for incubation with ≈1000 ng mL⁻¹ ApoE solution.

Plotting the Δi versus ApoE antigen amounts gives a straight line (R = 0.9989) for a large dynamic range (i.e. 0.13 to 1880 ng mL⁻¹) with RSD = 4.3%, which confirms the good reproducibility of the measurements (Fig. 3D).²⁴ Based on the correlation graph, the sensitivity of the immunosensor was estimated to be ca. 4.37 nA mL ng⁻¹. Chronoamperometry was also employed to corroborate the DPV findings. A series of chronoamperograms was recorded at 0.31 V, corresponding to the ferrocene oxidation potential with an ApoE concentration range between 0.47 and 1880 ng mL⁻¹ (inset of Fig. 3C). In agreement with the DPV data, the current response of ferrocene oxidation increased linearly with ApoE concentration.

3.3. Optimization of ferrocene chain length and chain anchor

To optimize the chain length of the ferrocene transducer, which is immobilized on the immunosensor surface, a series of different immunosensors using ferrocene derivatives with different chain lengths (i.e. FcL (6C) and FC_nS, with n = 2, 4, 8) were evaluated.

DPV studies (cf. Fig. S3A, S4A and S5A†) revealed that the highest sensitivity of the response of FC_nS-containing biosensors was obtained when the longest carbon chain (i.e. FC₈S) was utilized, reaching the value of ca. 4.26 nA mL ng⁻¹, which dropped by more than 57% for the shortest chain (Fig. 4A). These results were also confirmed by chronoamperometry (cf. Fig. S3B, S4B and S5B of ESI†), revealing a direct relationship between the chain length and the sensing performances (Fig. S6 of ESI†).

Fig. 4 (A) Increase of DPV current vs. ApoE amounts for the immunosensors prepared using ferrocene derivatives with different chain lengths: FC₈S (green; 8 carbon atoms), FC₄S (blue; 4 carbon atoms) and FC₂S (red; 2 carbon atoms) and corresponding fitting curves; (B) variation in DPV current vs. ApoE amounts for ApoE immunosensors functionalized with FC₈S (blue) and FcL (red) and correspondent fitting curves.

One plausible explanation for these observations is that the ApoE (34 kDa)²⁵ addition leads to the formation of a heavy immunocomplex, which will bend the alkyl chain and bring the ferrocene moiety closer to the surface.

Furthermore, the importance of the nature of the anchor site (i.e. the linker between AuNPs and ferrocene moiety) was also evaluated. Ferrocenyl-8-mercaptooctyl acid ester (FC₈S) and ferrocenyl lipoic acid ester (FcL) were selected for this study. The results shown in Fig. 4B suggest that the use of a disulfide anchor group leads to a more robust and sensitive biosensor, thanks to its two sulfur anchoring groups,^23,26 compared to the one with a single thiol site. In fact, based on the sensitivity of the electrochemical response of these two ApoE immunosensors, the sensitivity of FcL was found to be ca. 5.15 nA mL ng⁻¹ against ca. 4.26 nA mL ng⁻¹ for FC₈S.

3.4. Performance of the ApoE immunosensor

To evaluate the selectivity and the specificity of the ApoE immunosensor, different interfering proteins such as hIgG, gIgG, BSA and OVA were investigated to examine their effects on the immunocomplex formation. Selectivity tests were conducted by measuring the electrochemical response of the immunosensor to ca. 100 pg [94 ng mL⁻¹] of ApoE antigen and comparing it to that obtained in the presence of interfering proteins. For instance, the incubation of the immunosensor in the 100 ng mL⁻¹ solution of hIgG revealed no effect on the electrochemical response of the immunosensor, whereas the addition of ca. 100 pg of ApoE induces approximately 25% increase in current, as shown in DPV measurements (Fig. 5A). Similar results were obtained with BSA and gIgG (data not shown). Based on the results summarized in the histogram in Fig. 5B, we conclude that the immunosensor is highly selective to ApoE antigen.

Fig. 5 (A) DPV measurements for selectivity and specificity tests of the ApoE immunosensor in the presence of hIgG, BSA and gIgG as interfering proteins; (B) selectivity for ApoE; (C) histogram summarizing the current response observed for the selectivity test DPVs performed during the specificity study; (D) summary of results for the specificity test.

Additional experiments were performed to ascertain the ApoE immunosensor specificity. The results are summarized in the histogram presented in Fig. 5D. The DPV currents observed as a response to the antigen alone and to the antigen in the presence of a one-thousand-fold interfering protein (Fig. 5C) show no significant difference (i.e. less than 1.1%), which confirms the specificity toward ApoE. Thus, the prepared biosensor can discriminate ApoE in the presence of a large excess of hIgG, BSA, OVA and gIgG. Each measurement was run in triplicate to ensure the reproducibility of the response.

Since the levels of ApoE in human serum are in the μg per millilitre range,^14,15 the sample was diluted 1 : 5000 in order to meet the linear range of the reported ApoE immunosensor. The ApoE concentrations were measured and the obtained data are summarized in Table 1. Subsequently, 10, 20 and 30 ng mL⁻¹ of ApoE protein were injected into the solution and the final concentrations were measured a second time (Fig. S7†). The resulting high recovery values 105.1–109.8% with a standard deviation of less than 4.7% indicate the good performance of the developed ApoE immunosensor.

Table 1 Results of the detection of ApoE in real samples of human blood serum and recovery percentages after successive additions

Detected concentration (P)	Added concentrations (Q)	Detected concentrations after addition (R)	% recovery 100 × (R − P)/Q (±std dev.)
58.0	10.0	69.8	105.1 ± 3.4
—	20.0	82.3	107.4 ± 4.2
—	30.0	93.7	109.8 ± 4.7

---

Furthermore, the as-obtained ApoE immunosensor results were compared with those of other immunoassay biosensing systems developed to monitor the levels of ApoE protein. Table 2 summarizes the detection limits, concentration ranges, analysis times and linear slopes of ApoE with various other detecting methods such as CV, square wave voltammetry (SWV), ELISA and SPR. The current method exhibits great potential for ApoE sensing with a rapid analysis time compared to the ELISA method. In addition, there is a wider linear concentration range and a lower limit of detection compared to other techniques. It also has a high sensitivity as expected from the slope value of the calibration curve.

Table 2 Summary of works reporting ApoE immunosensing systems (MBs: magnetic beads, CM: cellulose membrane, GCE: glassy carbon electrode, CA: chronoamperometry, HER: hydrogen evolution reaction, WOR: water oxidation reaction, QDs: quantum dots, IrO₂NPs: iridium oxide nanoparticles, ITO: indium-doped tin oxide, fracAu: fractal gold nanostructures, ALP: alkaline phosphatase, FL: fluorescence, NA: non-available, Bt: butin, Avi: avidin)

Designed biosensor	Range (ng mL^−1)	LoD (ng mL^−1)	Time of analysis (hours)	Detection technique	Ref.
CM/αApoE/ApoE/α-ApoE/ALP	50–1000	0.04	24 h (overnight)	Amperometric/p-AP oxidation	16c
ITO/fracAu/α-ApoE/ApoE/α-ApoE/HRP	1–1000	0.30	5 h	CV	25
MBs/α-ApoE/ApoE/α-ApoE/IrO2NPs	100–1000	68	15 h (overnight)	CA/WOR	27
MBs/αApoE/ApoE/αApoE-Bt/Avi-CdSe@ZnS QDs	10–200	12.5	14 h (overnight)	SWV/QDs detection	28
MBs/α-ApoE/ApoE/α-ApoE/AuNPs	0.1–12.5	0.080	14 h (overnight)	CA/HER	29
ELISA plate/αApoE/ApoE/αApoE-Bt/Avi-CdSe@ZnS QDs	—	0.062	16 h (overnight)	ELISA/FL	30
ELISA plate/ApoE/PBS/αApoE/gIgG/ALP	25–200	—	17 h (overnight)	ELISA	31
ELISA plate/α-ApoE/ApoE/α-ApoE/β-galactosidase	NA	NA	13 h (overnight)	ELISA	32
ELISA plate/α-ApoE/ApoE/α-ApoE/HRP	NA	NA	17 h (overnight)	ELISA	14
SPCE/AuNPs/FcL/α-ApoE/ApoE	0.47–1800	0.13	5 h	DPV	This work

---

3.5. Mechanism of the immunosensing response

We speculated that the aforementioned increases in ferrocene oxidation currents during immunoreaction could be ascribed to the antigen/antibody complex formation. Hypothetically, the immunocomplex exerts a mechanical pressure on the ferrocene transducer to induce conformational changes in the alkyl chain bearing the end-capping ferrocene group. As was pointed out earlier, the ET rate constants from the ferrocene groups covalently immobilized on conductive surfaces are strongly dependent on the surface coverage and the conformational changes.⁵ In this scenario, the alkyl-terminated ferrocene moiety will bend to bring the ferrocene closer to the surface, which will allow the electron to hop, which in turn, will cause current increases, as illustrated in Fig. 6. In contrast, if the electrons travel to the electrode through a tunnelling mechanism, the ET rate should drop with the chain length increase.

Fig. 6 (A) As-proposed mechanism explaining the current increase during the immunocomplex formation for the two biosensors having long and short ferrocene linkers to gold, and (B) the variation plot of the difference, E_p − E_1/2, as a function of the logarithm of scan rate. Inset: cyclic voltammograms of the FC₈S-based immunosensor in the presence of 94 ng mL⁻¹ of ApoE.

To demonstrate this hypothesis, the series of ApoE immunosensors prepared previously were evaluated. The rate constants (k_ET) for ET between the redox centres and the surface were determined using the Laviron approximation based on the Butler–Volmer theory and the experimental measurements of the overpotential.‡^4,33 CV curves of the immunosensors in the presence of different ApoE concentrations were recorded at scan rates ranging from 200 to 4000 mV s⁻¹ (inset of Fig. 6B). The extracted values of k_ET from the plots of ΔE_p = E_1/2 − E_p (E_p is the anodic or cathodic peak potential and E_1/2 = (E_pa + E_pc)/2) as a function of the logarithm of scan rates (Fig. 6B) are summarized in Table 3. From this table, one can easily notice that the k_ET increases with increasing alkyl chain length bearing the ferrocene, as well as with the analyte concentration. For instance, the more flexible chains show larger increases in k_ET as compared to the shorter ones (i.e. 0.63 × 10³ s⁻¹ vs. 0.45 × 10³ s⁻¹ and 0.18 × 10³ s⁻¹ from C₈ to C₂, C₈ to C₄ and C₄ to C₂, respectively), as estimated when 188 ng mL⁻¹ of the analyte were added to the biosensors. These facts suggest that the chain flexibility and the mechanical strain exerted by the antigen govern the ET hopping from the ferrocene to the electrode surface. These findings support our proposed mechanism of the immunosensing response, since the aforementioned results regarding the improvement of ET rate constants with increasing ferrocenyl alkyl chain length differ from the case of ET by the tunnelling mechanism, wherein k_ET should proportionally decrease with the increasing chain length.⁴

Table 3 Values of k_ET determined for the alkyl-terminated ferrocene biosensors in the presence of different concentrations of ApoE

[ApoE4]/ng mL^−1	kET (FC2S)/s^−1	kET (FC4S)/s^−1	kET (FC8S)/s^−1
0	3.64 × 10^3	3.75 × 10^3	3.79 × 10^3
94	3.71 × 10^3	3.84 × 10^3	4.16 × 10^3
188	4.13 × 10^3	4.31 × 10^3	4.76 × 10^3

---

4. Conclusion

In this study, we presented a novel platform for the direct immunosensing of the ApoE protein. The label-free platform uses ferrocene covalently immobilized on α-ApoE/cysAuNPs on SPCE. The novelty in the concept is that the immunosensing event induces conformational changes in the alkyl chain, which increases the oxidation current of ferrocene through electron hopping from the ferrocene moiety to the electrode surface. The proposed mechanism was assessed by measuring the effect of the chain length and the ET rate constants on the immunosensor performances. The ApoE sensor exhibits a large dynamic range with a high sensitivity and a good detection limit of ca. 0.13 ng mL⁻¹.

Acknowledgements

The University of Tunis El-Manar and the Laboratory of Analytical Chemistry and Electrochemistry (LR99ES15) are gratefully acknowledged for the financial support and the mobility grants awarded to AMM and WA through the “Bourses d'Alternance” program.

Notes and references

1. D. Tang, B. Su, J. Tang, J. Ren and G. Chen, Anal. Chem., 2010, 82, 1527; N. Ronkainen, J. Thomas, B. Halsall and W. Heineman, Trends Anal. Chem., 2002, 21, 213; M. Dequaire, C. Degrand and B. Limoges, Anal. Chem., 2000, 72, 5521; L. Nyholm, Analyst, 2005, 130, 599; S. Arya, S. Singh and B. Malhotra, Handbook of Biosensors and Biochips, Wiley & Sons, Chichester, 2007.
2. D. Lu, J. Wang, L. Wang, D. Du, C. Timchalk, R. Barry and Y. Lin, Adv. Funct. Mater., 2011, 21, 4371; S. Choi, M. Goryll, L. Sin, P. Wong and J. Chae, Microfluid. Nanofluid., 2011, 10, 231.
3. G. Wang, X. Gang, X. Zhou, G. Zhang, H. Huang, X. Zhang and L. Wang, Talanta, 2013, 103, 75; H. Li, Q. Wei, J. He, T. Li, Y. Zhao, Y. Cai, B. Du, Z. Qian and M. Yang, Biosens. Bioelectron., 2011, 26, 3590; S. Kwon, H. Yang, K. Jo and J. Kwak, Analyst, 2008, 133, 1599; S. Dong, B. Wang and B. Liu, Biosens. Bioelectron., 1992, 7, 215; A. Garcia, C. Peniche-Covas, B. Chico, B. Simpson and R. Villalonga, Macromol. Biosci., 2007, 7, 435.
4. A. Eckermann, D. Feld, J. Shaw and T. Meade, Coord. Chem. Rev., 2010, 254, 1769.
5. R. Ruther, Q. Cui and R. Hamers, J. Am. Chem. Soc., 2013, 135, 5751.
6. K. Takeda, T. Morita and S. Kimura, J. Phys. Chem. B, 2008, 112, 12840.
7. J. Warren, M. Ener, A. Vlcek, J. Winkler and H. Gray, Coord. Chem. Rev., 2012, 25, 2478.
8. C. Fan, K. W. Plaxco and A. Heeger, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 9134; F. Ricci, A. J. Bonham, A. C. Mason, N. O. Reich and K. W. Plaxco, Anal. Chem., 2009, 81, 1608; K. J. Cash, F. Ricci and K. W. Plaxco, J. Am. Chem. Soc., 2009, 131, 6955; K. J. Cash, F. Ricci and K. W. Plaxco, Chem. Commun., 2009, 6222; A. Patterson, F. Caprio, A. Vallée-Bélisle, D. Moscone, K. W. Plaxco, G. Palleschi and F. Ricci, Anal. Chem., 2010, 82, 9109.
9. W. Lai, J. Zhuang, J. Tang, G. Chen and D. Tang, Microchim. Acta, 2012, 178, 357.
10. S. Viswanathan, C. Rani, A. Anand and J. Ho, Biosens. Bioelectron., 2009, 24, 1984.
11. Y. Ding, D. Li, B. Li, K. Zhao, W. Du, J. Zheng and M. Yang, Biosens. Bioelectron., 2013, 48, 281.
12. X. Han, Cell. Mol. Life Sci., 2004, 61, 1896.
13. M. Takeda, R. Martrinez, T. Kudo, T. Tanaka, M. Okochi, S. Tagami, T. Morihara, R. Hashimoto and R. Cacabelos, Psychiatry Clin. Neurosci., 2010, 64, 592; J. Kim, J. Basak and D. Holtzman, Neuron, 2009, 63, 287.
14. M. Lindh, M. Blomberg, M. Jensen, H. Basun, L. Lannfelt, B. Engvall, H. Sharnagel, W. Marz, L. Wahlund and R. Cowburn, Neurosci. Lett., 1997, 229, 85.
15. H. D. Soares, W. Z. Potter, E. Pickering, M. Kuhn, F. W. Immermann, D. M. Shera, M. Ferm, R. A. Dean, A. G. Simon, F. Swenson, G. A. Siuciak, G. Kaplow, M. Thambisetty, P. Zagouras, W. G. Koroshetz, H. I. Wan, G. Q. Trojanowski and L. M. Shaw, Arch. Neurol., 2012, 69, 1310; C. Wang, J.-T. Yu, H.-F. Wang, T. Jiang, C.-C. Tan, X.-F. Ming and H. D. Soares, PLoS One, 2014, 9, e89041.
16. (a) M. Porsch-Ozeurumez, S. Westphal and C. Luley, Clin. Chem., 2001, 47, 594; (b) K. Gaus and A. Hall, Anal. Chem., 1999, 71, 2459; (c) M. Meusel, R. Renneberg and F. Spencer, Biosens. Bioelectron., 1995, 10, 577.
17. A. Mars, C. Parolo, N. Raouafi, K. Boujlel and A. Merkoçi, J. Mater. Chem. B, 2013, 1, 2951.
18. S. Svarovsky, M. Gonzalez-Moa, M. Robida, A. Borovkov and K. Sykes, Mol. Pharm., 2009, 6, 1927.
19. W. Argoubi, M. Saadaoui, S. Ben Aoun and N. Raouafi, Beilstein J. Nanotechnol., 2015, 6, 1840.
20. M. Alonso, C. Yardimci, O. Renedo and M. Martinez, Anal. Chim. Acta, 2009, 633, 51.
21. J. Qui, H. Huang and R. Liang, Microchim. Acta, 2011, 174, 97; G. Jie, B. Liu, H. Pan, J. Zhu and H. Chen, Anal. Chem., 2007, 79, 5574.
22. I. Suni, Trends Anal. Chem., 2008, 27, 604.
23. R. Sahli, C. Fave, N. Raouafi, K. Boujlel, B. Limoges and B. Schollhörn, Langmuir, 2013, 29, 5360.
24. P. de Bièvre and H. Günzler, Validation in Chemical Measurements, Springer, Berlin, 2005.
25. Y. Liu, L.-P. Xu, S. Wang, W. Yang, Y. Wen and X. Zhang, Biosens. Bioelectron., 2015, 71, 396.
26. S. Chon and W. Paik, Phys. Chem. Chem. Phys., 2001, 13, 3405; W. Paik, S. Eu, K. Lee, S. Chon and M. Kim, Langmuir, 2000, 16, 10198.
27. L. Rivas, A. de la Escosura-Muñiz, J. Pons and A. Merkoçi, Electroanalysis, 2014, 26, 1287.
28. M. Medina-Sanchez, S. Miserere, E. Morales and A. Merkoçi, Biosens. Bioelectron., 2014, 54, 279.
29. A. de la Escosura-Muñiz, Z. Plichta, D. Horak and A. Merkoçi, Biosens. Bioelectron., 2015, 67, 162.
30. E. Morales-Narváez, H. Montón, A. Fomicheva and A. Merkoçi, Anal. Chem., 2012, 84, 6821.
31. C. Hesse, H. Larsson, P. Fredman, L. Minthon, N. Andreassen, P. Davidsson and K. Blennow, Neurochem. Res., 2000, 25, 511.
32. H. Song, K. Saito, M. Seishima, A. Noma, K. Unahami and K. Nakashima, Neurosci. Lett., 1997, 23, 175.
33. E. Laviron, J. Electroanal. Chem. Interfacial Electrochem., 1979, 101, 19–28; G. Lenaz and R. Fato, J. Bioenerg. Biomembr., 1986, 18, 369.

---

Footnotes

† Electronic supplementary information (ESI) available: Organic synthesis of the molecules, electrochemical measurements. See DOI: 10.1039/c5ra19328a

‡ Detailed equations for the determination of k_s using Laviron method's and others can be found in the review by Eckermann et al. (ref. 4).

---