Step‐by‐step monitoring of a magnetic and SERS‐active immunosensor assembly for purification and detection of tau protein

Abstract

We report a bottom‐up synthesis of iron oxide and gold nanoparticles, which are functionalized and combined to form a nanohybrid serving as an immune sensor, which selectively binds to tau protein, a biomarker for diagnosis of Alzheimer's disease. Detection of the target analyte is achieved by surface‐enhanced Raman scattering originating from the diagnostic part of the nanohybrid that was prepared from Au nanoparticles functionalized with 5,5′‐dithiobis‐(2‐nitrobenzoic acid) as a Raman reporter and monoclonal anti‐tau antibody. The magnetic part consists of Fe_xO_y nanoparticles functionalized with polyclonal anti‐tau antibody and is capable to separate tau protein from a complex matrix such as cerebrospinal fluid. We further identified and validated a set of analytical tools that allow monitoring the success of both nanoparticle preparation and each functionalization step performed during the assembly of the two binding sites by an immune reaction. By applying UV/Vis spectroscopy, dynamic light scattering, zeta potential measurements, X‐ray diffraction, small‐angle X‐ray scattering, and transmission electron microscopy, we demonstrate a proof‐of‐concept for a controlled and step‐by‐step traceable synthesis of a tau protein‐specific immune sensor.

A comprehensive characterization concept for a SERS‐based magnetic immunosensor with simplified preparation is reported here. Iron oxide and gold nanoparticles were functionalized with anti‐tau antibodies and used in a sandwich assay to separate tau protein from a biological matrix and enable its detection using SERS. Each individual step of the assembly is monitored by several methods to achieve a traceable and defined sensor formation.

Abbreviations

AD

Alzheimer's disease

CSF

cerebrospinal fluid

DLS

dynamic light scattering

DTNB

5,5′‐dithiobis‐(2‐nitrobenzoic acid)

EA

elemental analysis

EDC

1‐ethyl‐3‐(3‐dimethyl‐aminopropyl)carbodiimide

Fe_xO_y

iron oxide

HEPES

4‐(2‐hydroxyethyl)‐1‐piperazineethanesulfonic acid

IVDs

in‐vitro diagnostic devices

NHS

N‐hydroxysuccinimide

NP(s)

nanoparticle(s)

PEG

polyethylene glycol

SAXS

small‐angle X‐ray scattering

SERS

surface‐enhanced Raman scattering

SI

international system of units

TEM

transmission electron microscopy

TGA

thermogravimetric analysis

TNB

5‐thio‐2‐nitrobenzoic acid

XRD

X‐ray diffraction

1. INTRODUCTION

Over the last 15 years, nanoparticles (NPs) have found numerous applications in biotechnology and medicine 1, 2, 3, 4. For example, the plasmonic properties of noble metal NPs are used for imaging applications such as Raman spectroscopy to investigate molecules, pathogens, and tissues 5, 6, 7. In cases where the target analyte does not provide a sufficient Raman signal, plasmonic materials such as gold (Au) NPs can enhance the Raman‐cross sections of the adsorbed molecules, known as surface‐enhanced Raman spectroscopy (SERS) 8. If the target analyte does not directly interact with plasmonic NPs or does not itself provide a significant SERS spectrum, so‐called SERS tags may instead serve as sensitive analytical probes. A SERS tag consists of plasmonic NPs and a Raman‐active reporter molecule on the NP surface having binding sites selective toward the target analyte. Successful binding of the analyte is then detected indirectly through the Raman signature of the reporter 9, 10. Using suitable conjugation techniques, SERS tags can be further modified to provide additional functionalities, for example, through coupling of Au NPs to magnetic components such as iron oxide (Fe_xO_y) NPs. In this way, sensitive detection resulting from the plasmonic surface can be effectively combined with the ability for magnetic separation of the analyte from complex matrices 11, 12.

The ability of a SERS assay to magnetically capture the target analyte and separate it from potentially interfering matrix compounds is particularly important for the detection of disease‐specific biomarkers in body fluids such as cerebrospinal fluid (CSF). Several approaches to build up multicomponent SERS assays have been proposed for the identification and quantification of biomarkers such as enzymes 13, mRNA 14, 15, or viral DNA 16. A good overview of SERS‐based immunoassays has been recently presented by Lai et al 17. The high sensitivity of the SERS tag makes this method a promising approach for the detection and monitoring of diseases associated with exceptionally low biomarker concentrations. This applies, for example, to tau protein which is considered as one of the most important biomarkers for Alzheimer's disease (AD) 18, 19. Zengin et al. established such a multicomponent SERS tag for the quantification of tau protein with a limit of detection below 25 fmol/L 20.

In clinical chemistry and laboratory medicine, those markers are typically measured with the help of in vitro diagnostic devices (IVDs). Often a quality assurance of such routine analyses is required by law 21. For this purpose, so‐called reference measurement methods must be provided, which are ideally connected directly to the international system of units (SI) 21, 22, 23. It has already been demonstrated that SERS‐based methods are basically suitable for realizing the required SI‐traceability 24, 25. However, the establishment of a reference measurement method to quantify extremely low concentrations of large biomarker molecules such as tau protein is still challenging and time consuming. One of the reasons is that metrological validation requires the measurement result being associated with a measurement uncertainty. However, a reliable estimation of the uncertainty budget according to internationally accepted guidelines is only possible if the underlying process is fully understood and traceable 26.

Here, we present a defined route for the assembly of the entire SERS assay with a focus on monitoring each individual step of the preparation (Figure 1). In comparison to the system of Zengin et al. 20, a simplification of the magnetic component significantly reduced the synthesis duration and workload for preparation of the whole tau protein‐specific immunosensor. The proposed approach can also be transferred to other protein biomarkers by exchanging the specific antibody. After functionalizing Au NPs with 5,5′‐dithiobis‐(2‐nitrobenzoic acid) (DTNB), the Fe_xO_y‐ and Au NPs were modified with polyclonal and monoclonal anti‐tau antibodies, respectively. Finally, the complete SERS tag is formed by coupling of the two NP entities through the tau protein. The preparation of the particles as well as each step of the hybrid system formation were characterized by several methods, such as dynamic light scattering (DLS), elemental analysis (EA), small‐angle X‐ray scattering (SAXS), transmission electron microscopy (TEM), thermogravimetric analysis (TGA), UV/Vis spectroscopy, X‐ray diffraction (XRD), and zeta potential measurements.

Figure 1.

Schematic depiction of the hybrid complex formation with the individual experimental steps tagged and described below

2. MATERIALS AND METHODS/EXPERIMENTAL

2.1. Chemicals and reagents

The antibodies (monoclonal and polyclonal SP70 anti‐tau antibody, both produced in rabbit), 1‐ethyl‐3‐(3‐dimethyl‐aminopropyl)carbodiimide (EDC, ≥ 98%), gold‐(III) chloride (99%), 4‐(2‐hydroxyethyl)‐1‐piperazineethanesulfonic acid (HEPES, ≥ 99.5%), hydrochloric acid (37%), sodium‐(L) ascorbate (≥ 98%), N‐hydroxysuccinimide (NHS, ≥ 98%), sodium hydroxide (NaOH, ≥ 97%), polyethylene glycol 600 (PEG, techn. grade) and tau protein (Tau‐411 human expressed in E. coli, ≥ 90%) were obtained from Sigma‐Aldrich Chemie GmbH (Taufkirchen, Germany). DTNB (≥ 99%) was acquired from AppliChem GmbH (Darmstadt, Germany). Absolute ethanol (EtOH, 99.5%) was supplied by Fisher Scientific GmbH (Schwerte, Germany). Sodium citrate dihydrate (≥ 99.95%) was purchased from Honeywell Fluka (Seelze, Germany). Acetone (techn. grade) was obtained from VWR International GmbH (Darmstadt, Germany). The Fe_xO_y NPs were purchased as a powder from IoLiTec Inc. (Heilbronn, Germany). All chemicals were used as received.

To minimize the standard uncertainties of the concentrations, all solutions were handled gravimetrically. Unless otherwise stated, ultrapure water (18 MΩ·cm) was used, which was prepared from pre‐cleaned tap water using a Millipore Elix 5 UV device in combination with a Milli‐Q Element A10. The HEPES buffer was prepared by dissolving 0.063 g of HEPES powder in 50 g of ultrapure water and adjusting to a pH of 7.48–7.52 using a 0.1 mol/L NaOH solution.

2.2. Preparation of citrate‐modified iron oxide nanoparticles (a)

For the magnetic component, 0.17 g of Fe_xO_y powder was dispersed in 50 mL of sodium citrate solution (0.018 mol/L) and mixed in a 50 mL of stainless steel container with 54.3 g of 200 μm zirconia grindings 27. After 1 hour at 400 rpm in a planetary ball mill (PM400, Retsch GmbH), the suspension color turned black. Subsequent filtration of the solution through a 50 μm mesh separates the grinding media and after centrifugation at 1000 g for 5 minutes, sedimented bigger aggregates were removed from the solution and the supernatant with the remaining particles was mixed with acetone (5:1 v/v acetone/supernatant). The precipitated particles were centrifuged at 2400 g for 5 minutes and resuspended in ultrapure water after discarding the supernatant. This washing procedure was repeated for two more times. The resulting concentration of Fe_xO_y NP was about 3 mg/mL, as determined by quantification with phenanthroline 28. For further experiments, this solution was diluted with HEPES buffer in a ratio of 1:8 (v/v) and centrifuged at 10000 g for 10 minutes. After magnetic separation on a magnetic rack (MagRack 6, GE Heathcare), the supernatant was exchanged with HEPES buffer. Centrifugation and buffer exchange were performed twice.

2.3. Immobilization of anti‐tau antibodies on iron oxide nanoparticles (b)

To couple anti‐tau antibodies to the citrate modified Fe_xO_y NPs, 14 μL of an aqueous EDC solution (1.92 mg/mL) and 2 μL of an NHS solution (1.15 mg/mL) were added quickly one by one to 1 mL of Fe_xO_y NP solution. After incubation for 12 minutes, 2 μL of polyclonal anti‐tau antibodies were added, mixed carefully, and incubated at room temperature for 3 hours. Afterward, the particles were centrifuged for 20 minutes at 2000 g and after magnetic separation the supernatant was replaced with HEPES buffer, which was repeated twice.

2.4. Synthesis of gold nanoparticles (c)

The synthesis of Au NPs was carried out analogously to the patent of Taniuchi et al. 29 with some modifications. For the nucleation step (core colloid), gold‐(III) chloride (0.17 g) was dissolved in 25 mL of ultrapure water and sodium citrate dihydrate (0.49 g) was dissolved in 100 mL of ultrapure water. Six milliliters of the gold chloride solution (4.1∙10⁻⁴ mol/L) were diluted in 200 mL of ultrapure water, refluxed with heating in a PEG bath and stirred (ca. 700 rpm) for 30 minutes. After reaching a constant temperature of about 100°C, 50 mL of citric acid solution (1.6∙10⁻⁴ mol/L) was added, refluxed with heating, and stirred for another 15 minutes. Afterward, the Au NPs were cooled down to room temperature. For the synthesis of the final colloid, 233 mL each of a gold‐(III) chloride solution (1.7∙10⁻⁴ mol/L) and a sodium‐(L) ascorbate solution (4.7∙10⁻⁴ mol/L) were added (volumetric flow rate = 2 mL/min using a calibrated peristaltic pump) to 12 mL of the solution of core colloid (3∙10⁻⁴ mol/L, previously heated to 30°C) and stirred at 700 rpm. After complete addition of the precursors, the final solution was kept at 30°C for 1 hour, followed by centrifuging the particles (core colloids at 17000 g for 30 minutes, final colloids at 2000 g for 20 minutes) and replacing the supernatant with HEPES buffer (adjusted to optical density (OD) 1).

2.5. DTNB functionalization of gold nanoparticles (d)

Ten microliters of DTNB solution (10 mmol/kg in EtOH) were added to 1 mL Au NP dispersion. After 15 minutes incubation, the suspension was centrifuged for 20 minutes at 2000 g and the supernatant was replaced with HEPES buffer. The purification step was performed twice.

2.6. Immobilization of anti‐tau antibodies on DTNB‐functionalized gold nanoparticles (SERS tag) (e)

Two microliters of monoclonal anti‐tau antibody solution were added to a suspension of 1 mL DTNB‐functionalized Au NPs, which was incubated for 3 hours at room temperature. Afterward, the particles were washed with HEPES buffer by centrifuging the solution at 2000 g for 20 minutes and exchanging the supernatant for two times.

2.7. Capturing of tau protein with the SERS tag and separation with magnetic nanoparticles (f)

Two microliters of tau protein solution (1 μg/μL in ultrapure water) were added to 1 mL of the surface‐modified Au NPs and incubated at room temperature for 3 hours. Then, 1 mL of surface‐modified Fe_xO_y NPs were added and incubated over night at room temperature. The subsequent purification was a two‐fold centrifugation at 1000 g for 10 minutes, wherein a magnetic purification was carried out using a MagRack 6 for approximately 30 minutes. After the second washing step, the hybrid complex was diluted in 200 μL of HEPES buffer.

2.8. Characterization methods

DLS and zeta potential measurements were performed with a Zetasizer Nano from Malvern Instruments. For DLS, all samples were diluted by a factor of 10⁴–10⁵ to minimize occurring fluorescence and then measured three times with a 173° backscattering set‐up. Data evaluation was accomplished using the Zetasizer Nano software whereby the volume distributions were utilized for the assessment of hydrodynamic diameters. Zeta potentials were measured using a capillary zeta cuvette (DTS1070C, Malvern Panalytical Ltd).

EA was carried out with a FlashEA 1112 from Thermo Quest after drying the samples under vacuum conditions.

SAXS was performed with a SAXSess mc² from Anton Paar utilizing Cu‐K _α radiation (wavelength: 0.154 nm, voltage: 50 mA, current: 40 kV) and a CCD detector. At room temperature, a 1 mm flow‐through quartz cuvette was used at a sample‐to‐detector distance of 309 mm. Exposure times were between 0.1 and 100 seconds and the measuring times between 2 and 1000 seconds. The sample was measured in transmission mode with a slit collimation setup. The scattering data were routinely corrected 30 with respect to the background, transmission, instrumental noise, and smearing effects from the slit collimated beam. Basic steps were done in SAXSquant (Anton Paar GmbH, Austria), whereas desmearing and the indirect Fourier transformation (IFT) were carried out in GIFT (PCG software package, University Graz) 31. The IFT translates the scattered information into real space and yields a model‐free pair distance distribution function (P[r]), visualizing all frequencies of lengths between individual scatters within the system. The relatively most occurring distance is observable by the mode of the P(r), which equals the radius (R _max) for spherical particles.

SERS spectra were acquired using a LabRAM Aramis Raman microscope and LabSpec 5 Software from Horiba Jobin‐Yvon. The instrument was equipped with a 1200 grooves/mm holographic grating and a HeNe excitation laser (633 nm). For sample preparation, 200 μL of the nanohybrid solution were incubated on a silanized microscope slide overnight and were then air dried. Each sample was irradiated with an exposure time of 200 ∙ 1 seconds. The spot size was ~550 nm. The spectra were accumulated and measured three times each. Calibration of the spectral line position was carried out prior to each acquisition using the polystyrene ring breathing mode at 1001.4 cm⁻¹, according to the recommendations given in the ASTM guideline 32.

TEM images were taken with a Tecnai G² F20 TMP from Fei at 200 kV. The samples were applied onto a carbon film on a 3.05 mm woven copper net with 300 mesh from Plano GmbH.

TGA was performed using a TGA/DSC 1 STARe system and a gas controller 4C200 STARe system from Mettler Toledo. Fifteen milligrams of the dried sample were put in a ceramic crucible and measured at a heating rate of 10°C/min under an oxygen atmosphere.

UV/Vis spectra were recorded using a LAMBDA 1050 dual‐beam photometer from Perkin Elmer, measuring in a Quartz Suprasil semi‐micro cuvette (layer thickness 1.0 cm) from Hellma Analytics.

XRD was carried out with an Empyrean device from Panalytical. The measurements were carried out with Cu‐K _α radiation (wavelength λ of 0.154 nm) on a Si sample holder in a range of 2θ from 20 to 90° and a step size of 0.05°.

3. RESULTS AND DISCUSSION

3.1. Preparation of the magnetic component

After the dispersion process in the planetary ball mill, TEM images show Fe_xO_y NPs with sizes between 18 and 24 nm and divergent morphologies (Figure 2A). By increasing the focus, TEM images displaying the lattice fringes of the crystalline Fe_xO_y NPs could be obtained (Figure S1A). DLS measurements revealed a hydrodynamic diameter of approximately 24.7 nm with a polydispersity index (PDI) around 0.3 (Figure 2B). The tailing of the DLS curves toward larger particle sizes hints to the presence of some agglomerates, which is a classic result of top down processes since mechanical energy acts arbitrarily 33. A subsequent immobilization of anti‐tau resulted in an increase of the hydrodynamic diameter by approximately 10 nm (Figure 2B), about twice as much as was detected for other types of immunoglobulin G (IgG) with similar molar weight 34, indicating multiple binding on single Fe_xO_y NPs. UV/Vis measurements show an increase of the OD (Figure S1B) in the specific absorbance region of IgGs (approximately 250–280 nm, 35) after adding anti‐tau, which is most likely based on successful antibody immobilization.

Figure 2.

A, Representative TEM image of the citrate‐stabilized Fe_xO_y NPs, B, DLS results of Fe_xO_y NPs before and after antibody coupling with the respective median sizes and PDIs, C, XRD result of Fe_xO_y powder and a reference‐pattern of magnetite and maghemite, D, thermogram of Fe_xO_y powder before and after dispersion with citric acid

From the XRD measurements together with the reference patterns (magnetite (98‐015‐8714) and maghemite (98‐007‐9196) from the Inorganic Crystal Structure Database (ICSD)), a crystalline mixed phase can be assigned to the Fe_xO_y powder (Figure 2D). Due to the wide signal width, a distinction between the magnetite and maghemite phases is not possible. The higher baseline in the range between 25° and 35° indicates amorphous contents. By applying the Debye–Scherrer equation to the highest intensity reflection at 35.6° with an FWHM of 0.6° and K _S of 0.9, a crystallite size of 12.7 nm was calculated. The deviation of approximately 12 nm from the hydrodynamic diameter implies that the prepared Fe_xO_y NPs consist of several crystallites.

Using EA, an increase in the amount of carbon by 9.1% and hydrogen by 1.2% on the Fe_xO_y NPs was determined after dispersion treatment with added citrate, which is fully attributed to citrate bound to the particle surface. Considering the molar masses of citrate and Fe_xO_y NPs, a mass fraction of 26.6 wt% represents the stabilizing citrate layer. Additionally, TGA measurements of the Fe_xO_y powder show a loss of 2.0 wt% within the ramp to 150°C as a result of excess water desorption (Figure 2C). Upon further heating, an additional weight loss of 2.4 wt% due to thermal decomposition of adsorbed organic compounds is observed. In contrast, the citrate stabilized particles show a mass loss of 3.1 wt% due to evaporating water up to 185°C. A subsequent stepwise mass loss of altogether 31.1 wt% implies a multilayer system of citrate. The difference of the overall mass losses of 29.8 wt% (4.4 and 34.2 wt%) is the citrate contribution, which, considering the measurement accuracy, is in good agreement with the EA results.

3.2. Synthesis of the gold nanoparticles

A spherical morphology of both the core and final Au NPs was revealed by TEM with the core colloids having an average particle size of approximately 12.0 and 38.2 nm, respectively (Figure 3A). The TEM image of the final nanoparticles proves their relatively high uniformity. The difference between the hydrodynamic diameter (Figure 3B) and the SAXS‐derived diameter (calculated via R _max values, Figure 3C) is around 11.6 and 10.4 nm for the core colloid as well as 36.9 and 33.8 nm for the final colloid. This can be attributed to the measurement procedures, as for DLS measurements ligand effects exert a greater influence, whereas for SAXS only the crystalline core provides high electron contrast and contributes to the scattering. Furthermore, DLS measurements also indicate narrow size distributions and the SAXS results confirm spherical morphology as seen in the TEM images 36.

Figure 3.

A, TEM images, B, DLS measurements with the respective median sizes and PDIs, C, pair distance distribution functions with the respective R _max values and the corresponding diameters (in brackets), D, diffractogram with a gold reference‐pattern, and E, plots of the maximum OD (left) obtained via UV/Vis measurements and of the respective wavelength (right) as a function of the ripening time of the synthesized Au NPs

XRD measurements show the representative diffractogram of a synthesized Au NP system, with both the core colloid and final colloid reflections shown at almost same position, matching the reference pattern (98‐016‐3723) and differing only in their width (Figure 3D). A crystallite size of 31.4 nm for the final colloid was calculated by applying the Debye–Scherrer equation to the highest intensity reflection at 38.1° with a FWHM of 0.3°. The difference of the colloid size calculated via XRD, SAXS, and TEM is attributed to on an amorphous shell, which is also implied by the higher baseline in the range between 25° and 35°.

The long‐term stability of the colloids was investigated by UV/Vis measurements using the progress of the OD and position of the according maxima for up to 140 days after synthesis (Figure 3E). After 10 days, the OD and the maximum position of the core colloids remain constant at 519.0 nm. After about 30 days, the final colloids reach the plateaus around 523.6 nm. In the following 100 days, all observed factors remain constant. For both colloid sizes, the occurrence of Ostwald ripening is the explaining factor for the initial variations 37.

3.3. Hybrid complex assembly

After magnetic separation, a pinkish precipitate was obtained at the magnet‐facing tube wall, implying the presence of Au NPs bound to magnetic Fe_xO_y NPs (Figure 4A, left). After removing the unbound Au NPs, the solution still remained slightly pink (Figure 4A, right). Without tau protein or using another protein, a complete colorless solution was obtained after magnetic separation, showing the removal of all Au NPs (Figure 4A, right).

Figure 4.

A, Photographs of (left) magnetic separation of the hybrid complex, (right) resuspension after magnetic separation providing tau protein (magenta) and no or other proteins (colorless), B, TEM images of the synthesized hybrid NPs, C, zeta potential measurements of the NP systems for each process step, D, magnified UV/Vis spectra showing the absorption maxima of all individual Au NP systems, and E, normalized UV/Vis spectra of the synthesized hybrid NPs and their two colloidal precursors

The TEM images show that next to all round and high‐contrast Au NPs smaller Fe_xO_y NPs are located, which is attributed to the binding between the Fe_xO_y‐ and Au NPs via tau protein (Figure 4B, left). Furthermore, two large Au NPs are shown with smaller Fe_xO_y NPs in between which is either based on two different hybrid complexes with superposition of both complexes or on one big hybrid complex consisting of two Au NPs, as well as several Fe_xO_y NPs (Figure 4B, center). Different lattice fringes of crystalline Fe_xO_y and Au materials can be identified with a layer structure of the organic components DTNB and citrate enclosing both colloids (Figure 4B, right). In the Supporting Information, DLS spectra are presented in which an increase of the hydrodynamic diameter after each functionalization step can be observed (Figure S1C). Especially the modified diameter after addition of the functionalized Fe_xO_y NPs to the tau protein‐coupled Au NPs indicates a successful complex formation as well.

To evaluate each functionalization step as well as the formation of the complete hybrid NP system, the zeta potentials of the individual precursors and the final complex were measured (Figure 4C). Due to the presence of citrate at the NP surface, both NP systems initially show highly negative zeta potentials. After each surface functionalization step of the respective NP system, a shielding of surface charges and a higher mass and thus bigger volume occurs, which leads to a reduced electrophoretic mobility, hence leading to a decrease in the absolute value of the zeta potentials. Since the binding of the functionalized Au NPs to the functionalized Fe_xO_y NPs combines two massive nanoparticulate structures via tau protein, the change of the zeta potential to −14.1 mV is the most pronounced shift observed.

For further investigations, UV/Vis spectra were acquired after each functionalization step and normalized to the respective absorption maximum to unambiguously characterize the shift of the corresponding wavelength (Figure 4D). The occurring red‐shifts are based on an increased dielectric constant around the NP surface after a successful functionalization step 38. Mixing the functionalized Fe_xO_y NPs with the tau protein‐coupled Au NPs results in the largest shift by nearly 3.5 nm, possibly due to a major influence of the functionalized magnetic NPs. The full UV/Vis spectrum (250–700 nm) of the obtained hybrid system reflects both the Au‐specific absorption maximum (around 525 nm) and the strong Fe_xO_y‐specific increase of absorption from 450 nm toward lower wavelengths (Figure 4E). Since both the translocation of the maximum after each functionalization and the optical characteristics of both nanoparticulate components in the resulting UV/Vis spectrum can be detected, the evaluation of the UV/Vis spectra also clearly indicates the successful formation of the hybrid NP system.

3.4. Application of the hybrid complex

The washing procedures performed during the assembly of the hybrid complex always entail the risk of inducing precipitation of the Au NPs. Therefore, the number of washing steps should be kept to a minimum. The whole assembly process was monitored with Raman spectroscopy by evaluating the intensities of the Raman signals corresponding to the 5‐thio‐2‐nitrobenzoic acid (TNB) molecule (Figure 5A,B), which is bound to the Au NP surface after the disulfide group of DTNB has dissociated in solution 39, 40. Figure 5A shows the varying Raman intensities obtained during the preparation and washing steps while building up the Au NP moiety and the whole hybrid complex. It is obvious that excess and hence unbound antibody and tau protein quench the signal intensities in contrast to excess TNB. Due to their size, the protein as well as the antibody act as spacers which prevent the Au NPs from approximating each other to an extent required to achieve Raman signal enhancement. Additionally, after incubation with the antibody, only a broad background appears indicating a strong fluorescence, which completely suppresses the Raman spectrum. Washing the particles causes the signals to increase again as this treatment removes unbound substances and restores the associated ability to form plasmonic hotspots between the Au NPs.

Figure 5.

A, Variations of the DTNB SERS signal after various preparation steps in the course of the Au NP functionalization: AuNP + DTNB (1), after incubation (2) and subsequent washing step (3), after anti‐tau immobilization (4), incubation (5) and washing (6), after binding tau (7), incubation (8) and washing (9), B, SERS spectrum after hybrid NP synthesis in matrices with and without tau protein

To ensure that tau protein is the essential compound required to form one complex comprising Au‐ and Fe_xO_y NPs, a negative control was prepared and analyzed. Besides, as a proof of the applicability of the approach for analyzing real samples, measurements were carried out in CSF as the solvent (Figure 5B). It has been shown that in all cases where no tau protein has been added, that no TNB signals could be detected. This implies that no TNB‐covered Au NPs remain in the sample solution and that only the presence of tau protein causes TNB signals to appear in the spectra proving the specificity and the selectivity of the complex for tau protein.

4. CONCLUSIONS

In this study, a comprehensive characterization concept for a facile and defined formation of a SERS‐based magnetic immunosensor is reported. Magnetic Fe_xO_y NPs with a citrate multilayer and a hydrodynamic diameter of about 24.7 nm were prepared and successfully conjugated with polyclonal anti‐tau antibody. Spherical and highly uniform colloidal Au NPs with a final size of approximately 36.0 nm as well as a narrow particle size distribution were obtained in a two‐step preparation process followed by aging through Ostwald ripening for several weeks. Successful functionalization steps as well as the formation of the hybrid complex caused a decrease in the absolute value of the zeta potential, a larger hydrodynamic diameter as well as a stepwise red‐shift of the absorption maximum. These results confirm the successful synthesis of a hybrid system having a defined structure as confirmed by the TEM images. As a proof‐of‐concept, the formation of the hybrid complex in tau protein‐containing solutions enabled a detection of TNB spectra via SERS. Furthermore, the methods used can serve as a template for the characterization of magnetic SERS‐based immunosensors and are summarized in Table 1.

Table 1.

Characterization concept containing measuring methods, resulting particle properties, and their overall importance for the establishment of SERS‐based immunosensors

	Method	Measured properties	Importance
Primary particle synthesis	DLS	Agglomeration state	Required for quality control
	SAXS	Particle size, agglomeration state	Optional
	TEM	Particle size, morphology	Required for quality control
	TGA	Quantification of adsorbed organics	Optional if an established synthesis strategy is used; only useful for FexOy NPs
	UV/Vis	Changes in particle size (e.g., fluctuations of absorption due to Ostwald ripening)	Optional; only relevant for Au NPs
	X‐ray diffraction	Particle size, crystallinity	Recommended
	Zeta‐potential	Surface charge based on the electrophoretic mobility	Required for quality control
Stepwise functionalization	DLS	Increase of hydrodynamic diameter; possible agglomeration	Optional if an established strategy is used
	TEM	Structural integrity of complex	Recommended
	UV/Vis	Absorption profile and shift	Required for quality control
	Zeta‐Potential	Change of the electrophoretic mobility and surface charge	Required for quality control
Proof‐of‐concept	SERS	SERS‐signal after magnetic separation	Crucial step

The simplified preparation concept for a controlled formation for a SERS‐active immunosensor greatly improves the applicability of this approach and allows a larger scale implementation. Furthermore, a quantification principle will be established utilizing the isotope dilution (ID) technique, which is mostly used in mass spectrometry, but which has also been shown to work excellently when combined with Raman spectroscopy 24, 25, 41, 42, 43. Besides the advantages of having molecular specificity, accuracy and a high sensitivity during quantification, an ID‐SERS approach also could provide SI‐traceability of the measurement results when combined with well‐characterized reference materials. In this way, reference measurement procedures can be implemented and also transferred to other biomarkers just by substituting the used antibodies.

CONFLICTS OF INTEREST

The authors declare no potential conflict of interest.

Supporting information

Appendix S1: Supporting information

Maurer V, Frank C, Porsiel JC, Zellmer S, Garnweitner G, Stosch R. Step‐by‐step monitoring of a magnetic and SERS‐active immunosensor assembly for purification and detection of tau protein. J. Biophotonics. 2020;13:e201960090 10.1002/jbio.201960090