We present a study aimed at developing a label-free optical fiber biosensor for detection and quantification of biomolecules in real-time. The biosensor based on a Tilted Fiber Bragg Grating (TFBG) transduces a binding event between the probe and target molecules into a change in the refractive index of the medium surrounding the fiber. This work describes the experimental results obtained with three methods for immobilizing biomolecular probes on a TFBG silica cladding surface. Bovine serum albumin (BSA) and anti-BSA are used to assess the performances of the TFBG based biosensor in each configuration.
©2008 Optical Society of America
For three decades there has been a growing interest in the design of biosensors aimed at detection, diagnosis, and determination in the fields of food and water-quality control, health, safety, and environmental monitoring. A biosensor usually couples an immobilized bio-specific recognition molecule to the surface of a transducer, which converts a molecular recognition event into a measurable signal, pinpointing the presence of the target molecule. The signal generated is often directly proportional to the target molecule concentration. The concept of combining the recognition properties of biological molecules with the sensitivity of transducers such as optical sensors has led to the emergence of optical biosensors as valuable sensitive and selective tools in analytical chemistry.
Fiber optic evanescent wave biosensors have become increasingly popular as an analytical tool in biomedical, biochemical and environmental applications. The main advantage of these sensors, beyond the fact that it is a label-free method, is their ability to give rapid and sensitive detection of the target biomolecule in real time. Evanescent waves generated at the interface of two optically different transparent media are known to decay exponentially. Decay of these electromagnetic radiations has been characterized by the penetration depth, the distance from the interface at which the amplitude of the intensity would fall to 1/e of its value at the interface. Typically, penetration depth is a fraction of incident wavelength, for the chemical sensing in the aqueous medium. Owing to its limited range, an evanescent wave can interact selectively with the molecules at the interface, without interference from the free molecules in solution. This optical interaction forms the basis for the optical waveguide-based biosensor. An optical fiber diffraction grating is used in this work as a transducer for the detection of interactions between the medium surrounding the optical fiber and evanescent waves. Fiber optic long period gratings, which also act as core-cladding mode couplers, have already been exploited to develop biosensing applications  while the possibility of using plasmon resonances in gold-coated tilted fiber Bragg gratings (TFBG) to detect DNA target has been recently investigated . In the present paper, we report on what we believe to be the first study demonstrating the relevance of a biosensor based on a TFBG refractometer directly biofunctionalized onto the silica cladding surface.
One of the great challenges to be met in the preparation of biosensors is the strategy for immobilization of the biospecific molecule on to the transducer. This work presents an experimental comparison of three methods for immobilizing biomolecular probes on an optical fiber silica cladding surface. The biosensor based on a Tilted Fiber Bragg Grating (TFBG) refractometer enables to directly detect, in real-time, target molecules. So, bovine serum albumin (BSA) (antigen) and anti-BSA (antibody) are used to study the reaction kinetics of the antigen-antibody recognition by changing the antibody concentration in the three configurations employed for the antigen immobilisation. The experimental data are described by the Langmuir isotherm model.
2. Tilted Fiber Bragg Grating refractometer
2.1. Tilted FBG spectral sensitivity to the refractive index
A standard Fiber Bragg Grating (FBG) consists of a refractive index modulation in the core of an optical fiber that is perpendicular to the fiber axis. An FBG reflects one characteristic wavelength corresponding to the so-called Bragg wavelength and transmits all others. In a Tilted Fiber Bragg Grating (TFBG), the variation of the refractive index forms an angle with the optical axis in order to enhance coupling between the forward-propagating core mode and the counterpropagating cladding modes, observed as numerous resonances in the transmission spectrum.
The resonances composing the spectral transmission response of a TFBG are modified by external refractive index changes . Wavelength resonances associated with propagating modes which effective index is lower than the external refractive index are not guided any more, as the interface between the cladding and the external medium has disappeared. To determine the external refractive index changes, the apparent sinusoidal nature of the complex spectrum of the TFBG is exploited . The Fourier transform of the transmission spectrum is determined and we observe that there is a continuum of frequencies.
When the external refractive index increases, the frequencies involved in the FFT (Fast Fourier Transform) of the transmission spectrum gradually disappear. The explanation is as follow: when the external refractive index increases, the amplitudes of low wavelength resonances are reduced. The lower the wavelength resonance is, the more the spectral distance between two consecutive dips/peaks increases (Fig. 1). Therefore, low wavelength resonances contribute to the low frequency component on the FFT spectrum (Fig. 2).
This analysis shows that, when the external refractive index changes, the predominant frequency of the Fourier transform of the TFBG spectral response can be considered for sensing. So, we establish a calibration curve of frequency variations versus refractive index values in order to use such a device for universal refractive index sensing purposes.
2.2. Real time detection system
In this work, we use TFBGs with external tilt angle of 20° in order to be able to observe refractive index variations in solutions that have refractive indices close to that of water at wavelengths between 1490 and 1590 nm. In our set up, TFBGs are interrogated by coupling the emission of a spectrally tunable laser source of wavelength ranging from 1490 nm to 1590 nm into the optical fiber, and monitoring the transmission spectrum using an InGaAs photodiode. An entire transmission spectrum is saved every 5 seconds (Fig. 3).
The process of biofuntionalization results in variations in the refractive index close to the vicinity of the fiber surface, which are monitored with the TFBG-based transducer. Before the biofunctionalization, the TFBG is calibrated to convert measured frequencies into refractive index values. The calibration curve (Fig. 4) has been established by using a set of refractive index liquids (Cargille’s oils ) of which the indices are perfectly known, with an accuracy of 0.0002 refractive index unit (r.i.u.).
3. Biofuntionalization of Tilted Fiber Bragg Grating
Preparation of fiber optic biosensors requires immobilization of bioreceptors on the optical fiber silica surface. Reported methods for immobilizing bioreceptors include adsorption , ionic bonding by electrostatic self-assembly technique , cross-linking by means of a multi-functional reagent , covalent bonding [9, 10] or by means of an avidin-biotin linkage . Biofuntionalisation processes commonly require the interaction between biomolecules and optical fiber surface by means of a stable intermediate layer. This interlayer must provide adequate functional groups that react with biomolecules (Fig. 5).
The TFBG cladding surface used for protein immobilization in our experiment is biofunctionalized by means of three different methods: ionic bondings only, ionic bondings combined with an avidin-biotin linkage and covalent bondings combined with an avidin-biotin linkage. The probe protein used is the Bovine Serum Albumin (BSA). As the refractive index value depends on the surrounding solution, experimental values used to study binding interactions correspond to the refractive index measured for the biosensor immersed in the buffer solution and after the washes.
3.1. Ionic bonding method
3.1.1. Deposition of electrostatic self-assembled film
The Electrostatic Self-Assembled (ESA) film deposition method is based on the electrostatic attraction between oppositely charged molecules in each monolayer deposited, and involves several steps. First, the optical fiber is cleaned in a 1M HCl solution for 20 minutes and treated in a 1 M NaOH solution for 30 minutes to create a negatively charged surface. Then, the substrate is dipped alternatively into cationic polymer (polycation) and anionic polymer (polyanion) solutions in order to create polyelectrolyte multilayers (Fig. 6). The polyelectrolytes used in the fabrication of ESA films are poly-ethylene-imine (PEI) as polycation, and dextran sulfate sodium (DSS) as polyanion. We prepared 1 g/L PEI and DSS solutions in a 0.01 M phosphate buffer solution (PBS), pH 7.4. The exposure time to each polyelectrolyte solution is 7 minutes. Between each adsorption step the layer is washed with PBS (Fig. 7). In order to determine the most appropriate thickness of the polyelectrolyte multilayer film, refractive index variations were measured as a function of the number of deposited bilayers. We observed that the refractive index increases with the number of deposited bilayers but we notice a lower sensitivity with less than four and more than twelve bilayers. As our objective is to develop a biosensor characterized by a wide sensitivity range, the polyelectrolyte multilayer is composed of four bilayers with an external positively-charged PEI layer to ensure the immobilization of the negatively charged probe BSA protein .
3.1.2. Protein adsorption
Proteins are affected by pH of their surrounding environment and can become more positively or negatively charged. By definition, the Isoelectric Point (pI) of a protein stands for the pH at which positive and negative charges are equal. At this pH, the sum of charges, over all amino acids, is null. At a pH below their pI, proteins carry a net positive charge. Above their pI, they carry a net negative charge.
We prepare a 10 g/L BSA solution in a 0.01M Phosphate Buffer Solution (PBS), pH 7.4. As this pH is higher than the isoelectric point of BSA (4.7), the protein is negatively charged and can be adsorbed by electrostatic interactions to the fiber surface, which is positively charged (Fig. 5.a) as the last monolayer of ESA film is a PEI monolayer . Then, the BSA solution is incubated for 30 minutes.
As charges are distributed randomly along the protein chain, the orientation of BSA biomolecules cannot be controlled. Moreover, a self-assembled multilayer does generally not provide a homogeneous environment as required in order to obtain the homogeneity of the protein layer. Therefore, the formation of monolayer coverage of probe proteins effectively occurs only at a sufficiently high concentration of BSA. A strategy to control the protein arrangement and to improve biomolecule activity consists in using biotin-extravidin mediated immobilization.
3.2. Ionic bonding combined with avidin-biotin linkage method
The avidin-biotin immobilization technique offers a variety of advantages. First, it is an extremely specific and strong non-covalent binding method. Second, avidin provides a passivation layer over the substrate surface which subsequently helps to prevent non-specific adsorption of biomolecules on the surface. Third, the avidin-coated substrate can be homogeneously linked with biotinylated proteins. Finally, the avidin-biotin immobilization procedure generally maintains bioreceptor binding activity more successfully that other regularly used methods .
The avidin-biotin immobilization method requires, first, a process to attach the avidin to the transducer surface. The easiest method employed for optical fiber based biosensors is electrostatic adsorption on a self-assembled multilayer.
A 1 g/L extravidin solution in PBS is incubated for 30 minutes. As this pH is higher than the isoelectric point of extravidine (~5), the protein is negatively charged and can be adsorbed by electrostatic interactions to the positively charged last layer of the ESA film. As extravidin exhibits extremely high affinity for biotin and provides a homogeneous environment, a lower concentration of BSA is required for the same monolayer coverage as compared to the probe proteins adsorbed by electrostatic interactions. So, we prepare a 1 g/L biotinylated BSA solution and the extravidin-coated substrate is linked with biotinylated BSA protein probes via avidin-biotin interactions (Fig. 5.b).
This procedure, even if it has proved successful, has some disadvantages, particularly because electrostatic adsorption methods frequently suffer from low long term stability. A strategy to improve the long term stability consist of substituting the ESA film for polymer brushes covalently bonded on the TFBG cladding surface.
3.3. Covalent bonding combined with avidin-biotin linkage method
We apply a “grafting-from” polymerization scheme  to grow covalently bound polymer chains from the optical fiber surface with high grafting densities. We use polyacrylic-acid which is prepared by free radical polymerization from the TFGB-surface under UV irradiation at room temperature (Fig. 8).
To prepare the silica surface, the later is treated with 10 % nitric acid solution for 2 hours in order to increase the surface density of silanol groups and activate the substrate for the silanization procedure. The activated silica is silanized with 10% 3-aminopropyl-trimethoxysilane (APTS) in water at room temperature for 1 hour, washed with water, rinsed with ethanol and dried at 80°C under nitrogen atmosphere for 15 hours. This is followed by the covalent coupling of a free radical initiator with 0.25 M 4,4’-azobis(4-cyanovaleric acid) chloride  in dry dimethylformamide (DMF) solution at room temperature.
To perform free radical polymerization of acrylic-acid, the TFBG-surface is immersed in 5 % acrylic-acid. A radical is formed by irradiation of the chemical initiator grafted on the fiber surface. The creation of this initiation site is followed by a free radical polymerization of acrylic acid monomers because of the radical propagation. After polymerization, the newly formed film is washed with water (Fig. 9).
As covalent attachment regularly results in partial denaturation of the biomolecules and conformational changes leading to reduction of their activity, the resulting polyacrylic-acid (PAA) is used to graft extravidin instead of BSA.
Attachment to PAA is achieved by utilizing the reactivity of terminal COOH groups of the polymer. To couple the extravidine to the surface through the formation of a peptide bond, the terminal carboxyl groups on the PAA are converted into reactive N-hydroxysuccinimide (NHS) esters intermediate by using a solution of 0.2 M 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC), 0.05 M NHS in water for 15 minutes. Then, the substrate is dipped for 1 hour into a 1 g/L extravidine solution in PBS. Finally, the extravidin-coated substrate is homogeneously linked with biotinylated BSA protein probes, immobilized on the optical fiber cladding surface via avidin-biotin interactions (Fig. 5.c).
Equilibrium isotherms are obtained for each functionalization protocol in order to quantitatively evaluate the three TFBG-based biosensors. Experiments based on only ionic bonding and ionic bonding combined with avidin-biotin linkage and biofunctionalization methods have been repeated two times and those based on covalent bonding combined with avidin-biotin linkage biofunctionalization method had not been repeated yet.
The graphical expression of isotherm is a plot of refractive index shift against the anti-BSA concentration in the medium surrounding the biofunctionalized surface. For each value of anti-BSA concentration, the binding of anti-BSA to the immobilized BSA is observed in real time. The Langmuir isotherm model is applied to describe the antigen-antibody interaction in order to assess sensitivities and anti-BSA concentration limits of the three biosensors resulting from the different methods of biofunctionalization.
4.1. Real-time antibody detection experiments
The real-time antibody detection experiments resulted in variations in the refractive index in the vicinity of the fiber surface, which are monitored with the TFBG-based sensor. The biosensor signal reaches a plateau after an incubation time depending on the antibody concentration; the higher the antibody concentration, the shorter the incubation time. The incubations are performed with samples containing gradual concentrations of anti-BSA. Figure 10 shows the biosensor response in terms of refractive index variations with antibody concentration.
The higher the concentration of anti-BSA is, the more the refractive index difference between the buffer and the anti-BSA solution increases. As the refractive index value depends on the concentration of the surrounding medium, experimental values used to study the refractive index shift as a function of anti-BSA concentration correspond to the refractive index measured for the biosensor immersed in the buffer solution and after the washes. Furthermore, non-specific interactions are eliminated by rinsing.
4.2. Langmuir isotherm model
The study of the refractive index shift as a function of anti-BSA concentration is completed with the Langmuir isotherm model. Three essential premises of the Langmuir isotherm are monolayer coverage, adsorption site equivalence and independence. The Langmuir equation can be expressed as follows:
where Δn is the refractive index shift corresponding to the binding of the anti-BSA onto the BSA-immobilized surface, C the concentration of the anti-BSA in solution, Δnmax the refractive index shift corresponding to the saturation (for the complete monolayer) and K the affinity constant of the antibody-antigen recognition.
Dots in Fig. 11 show the experimental biosensor responses in terms of refractive index variations with antibody concentration in the three configurations described above.
The adsorption data are studied by a regression analysis to fit the Langmuir isotherm model . The values of affinity constants K and refractive index shifts corresponding to the saturation Δnmax are deduced (Tab. 1). Selected Langmuir isotherm plots are displayed in figure 11 (solid curves for the first data sets and dashed curves for the second data sets). These isotherm models are in good agreement with the experimental data and suggest that the anti-BSA proteins form a monolayer  at the interface between the biofunctionalized surface and the surrounding medium.
4.3. Sensitivity and detection limit
The sensitivity for anti-BSA detection is given by
where Δn max is the refractive index shift at saturation and σmax the surface density concentration of anti-BSA target biomolecules when all available binding sites are occupied . Analysis of the estimates shows that the Δn max values are almost the same in the three cases. As the saturation corresponds to an occupation of the total number of binding sites, this result indicates an equivalent surface density of available binding sites in the three configurations investigated. Thus, a theoretical surface density concentration of anti-BSA at saturation, which is linked with the average length of an anti-BSA molecule d=12 nm , can be calculated by assuming monolayer coverage of the outer surface of the fiber.
where M=147,000 g/mol is the molecular mass of anti-BSA and NA=6.02·1023 is the Avogadro’s number. σmax is common for each protocol.
The refractive index shift is directly proportional to the surface density concentration of anti-BSA target biomolecules and the proportionality coefficient is given by the sensitivity. Therefore, the detection limit can be calculated by
where R=10-5 is the TFBG refractometer resolution. Currently, we develop other data acquisition system which acquires the data in 0.005-second intervals (200 Hz) so that in the future we can reach ten times better resolution. As the Δn max values are almost the same in the three configurations, the anti-BSA detection sensitivities and detection limits are also the same for all three biofunctionalization processes (Tab. 2).
4.4. Affinity and concentration limit
The affinity constant K reflects the activity of the surface immobilized protein which is a key parameter to assess the biosensor efficiency. The low activity of protein results in decreased
antibody-antigen recognition on the surface. Therefore, the affinity will deteriorate dramatically if the protein undergoes denaturisation on the surface.
The K parameters obtained with the Langmuir model show that the affinity constant corresponding to the antibody-antigen recognition on the surface is affected by the process used to immobilize the BSA protein.
The anti-BSA concentration limit is the anti-BSA concentrationClim giving a refractive index shift equal to the TFBG refractometer resolution R=10-5. As refractive index shift is given by the equation 1, the anti-BSA concentration limit is
Therefore, higher affinity constants result in lower anti-BSA concentration limits (Tab. 3). This result shows that using biotin-avidin mediated immobilization improves the activity of the surface grafted BSA proteins. The methods 2 and 3 based on avidin-biotin linkage maintain higher affinity and optimize the concentration limit associated to the detection limit of the TFBG based biosensor. These results confirm the hypothesis that BSA can be distorted by the electrostatic interaction with the surface, leading to a lower affinity for anti-BSA and that using biotin-extravidin mediated immobilization is an efficient strategy to control the protein arrangement and to improve biomolecule activity. Furthermore, the concentration limit could be improved to 10 µg/L by using a newly-developed data acquisition system which provides data at a rate of about 200 Hz.
In the fast growing sector of biosensor technology, an optical technique based on Tilted Fiber Bragg Gratings has been found to be potentially suitable to biomolecule detection. By immobilizing one component of a biospecific pair (antibody-antigen), the technique can be used for detection and measurement of the concentration of the counterpart, or for characterization of the kinetics and affinity of binding, by monitoring the change in the refractive index during interaction.
In this work, we have studied the immobilization of the BSA showing that the immobilized proteins retain their biological activity and can be used to detect anti-BSA. More importantly, we confirmed that the biofunctionalization processes based on avidin-biotin linkage (the second and third options) result in improved biological activity and enable to reach TFBG detection limit for lower anti-BSA concentrations as compared with the first option.
Further experiments are needed to calculate mean values of Langmuir parameters with standard deviations and determine if significant differences exist between the methods 2 and 3. Nevertheless, this technique seems promising as it makes it possible to consider the realization of a new kind of optical biosensors for a large range of applications, including measurements for human diseases, process control for the food industry (milk, wine…) and inspection sensors for environmental pollution monitoring.
References and links
1. M. P. DeLisa, Z. Zhang, M. Shiloach, S. Pilevar, C. C. Davis, J. S. Sirkis, and W. E. Bentley, “Evanescent Wave Long-Period Fiber Bragg Grating as an Immobilized Antibody Biosensor,” Anal. Chem. 722895–2900 (2000). [CrossRef] [PubMed]
2. Y. Y. Shevchenko, D. A. Blair, M. C. Derosa, and J. Albert, “DNA Target Detection Using Gold-Coated Tilted Fiber Bragg Gratings in Aqueous Media,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science Conference and Photonic Applications Systems Technologies, OSA Technical Digest (CD) (Optical Society of America, 2008), paper CMJ4, (2008). http://www.opticsinfobase.org/abstract. cfm?URI=CLEO-2008-CMJ4.
3. M. C. Phan Huy, G. Laffont, V. Dewynter, P. Ferdinand, L. Labonté, D. Pagnoux, P. Roy, W. Blanc, and B. Dussardier, “Tilted Fiber Bragg Grating photowritten in microstructured optical fiber for improved refractive index measurement,” Opt. Express 14, 10359–10370 (2006). [CrossRef] [PubMed]
4. S. Maguis, G. Laffont, P. Ferdinand, M. C. Millot, K. Kham, and S. Péralta, “Tilted fibre Bragg gratings for the specific detection of biological species,” Proceedings of 19th Int. Conf. Optical Fiber Sensors Perth, Australia, April 14–18, 2008.
7. C. Elosúa, C. Bariáin, I. R. Matías, F. J. Arregui, A. Luquin, and M. Laguna, “Volatile alcoholic compounds fibre optic nanosensor,” Sens. Actuators B 115, 444–449 (2006). [CrossRef]
9. S. Löfås and B. Johnsson, “A novel hydrogel matrix on gold surfaces in surface plasmon resonance sensors for fast and efficient covalent immobilization of ligands,” J. Chem. Soc. 21, 1526–1528 (1990).
10. D. M. Disley, J. Blyth, D. C. Cullen, H. X. You, S. Eapen, and C. R. Lowe, “Covalent coupling of immunoglobulin G to a poly (vinyl) alcohol-poly (acrylic acid) graft polymer as a method for fabricating the interfacial-recognition layer of a surface plasmon resonance immunosensor,” Biosens. Bioelectron. 13, 383–396 (1998). [CrossRef] [PubMed]
11. R. S. Marks, A. Novoa, D. Thomassey, and S. Cosnier, “An innovative strategy for immobilization of receptor proteins on to an optical fiber by use of poly (pyrrole-biotin),” Anal. Bioanal. Chem. 374, 1056–1063 (2002). [CrossRef] [PubMed]
12. D. S. Salloum and J. B. Schlenoff, “Protein adsorption modalities on polyelectrolyte multilayers,” Biomacro-molecules 5, 1089–1096 (2004). [CrossRef]
14. B. Zhao and W. J. Brittain, “Polymer brushes: surface-immobilized macromolecules,” Prog. Polym. Sci. 25, 677–710 (2000). [CrossRef]
15. R. Schmidt, T. Zhao, J. B. Green, and D. J. Dyer, , “Photoinitiated polymerization of styrene from self-assembled monolayers on gold,” Langmuir 18, 1281–1287 (2002). [CrossRef]
17. S. Brunauer, L. S. Deming, W. E. Deming, and E. Teller, “On a Theory of the van der Waals Adsorption of Gases,” J. Am. Chem. Soc. 62, 1723–1732 (1940). [CrossRef]
18. C. A. Barrios, M. J. Bañuls, V. González-Pedro, K. B. Gylfason, B. Sánchez, A. Griol, A. Maquieira, H. Sohlstrőm, M. Holgado, and R. Casquel, “Label-free optical biosensing with slot-waveguides,” Opt. Lett. 33, 708–710 (2008). [CrossRef] [PubMed]
19. J. Yang, M. Mayer, J. K. Kriebel, P. Garstecki, and G. M. Whitesides, “Self-Assembled Aggregates of IgGs as Templates for the Growth of Clusters of Gold Nanoparticles,” Angew. Chem. Int. Ed 43, 1555–1558 (2004). [CrossRef]