Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Biosensors and Bioelectronics 33 (2012) 190–195 Contents lists available at SciVerse ScienceDirect Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios Immobilization of enzyme on long period grating fibers for sensitive glucose detection Akash Deep, Umesh Tiwari ∗ , Parveen Kumar, Vandana Mishra, Subhash C. Jain, Nahar Singh, Pawan Kapur, Lalit M. Bharadwaj Central Scientific Instruments Organisation, Sector (CSIR-CSIO) 30 C, Chandigarh, India a r t i c l e i n f o Article history: Received 15 September 2011 Received in revised form 14 December 2011 Accepted 28 December 2011 Available online 6 January 2012 Keywords: Long period grating Glucose oxidase FTIR spectroscopy Raman spectroscopy Biosensing Glucose a b s t r a c t Glucose oxidase (GOD) immobilized long period grating (LPG) fibers have been proposed for the specific and sensitive detection of glucose. The treatment of LPG fibers with aminopropyl triethoxysilane has induced biding sites for the subsequent GOD immobilization. Field emission scanning electron microscopy, confocal laser scanning microscopy, infrared spectroscopy and Raman spectroscopy have provided detailed evidences about the effectiveness of the adopted biofunctionalization methodology. The enzyme activity is conserved during the immobilization step. Fabricated LPG sensor was tested on different glucose solutions to record the transmission spectra on an optical spectrum analyzer. The wavelength shifts in the transmission spectra are linearly correlated with the glucose concentration in the range of 10–300 mg dL−1 . The fabricated sensor gives fast response and is demonstrated to be of practical utility by determining glucose contents in blood samples. Proposed technique can further be extended to develop LPG fiber based novel, sensitive and label free nanosensors for disease diagnosis and clinical analysis. © 2012 Elsevier B.V. All rights reserved. 1. Introduction Fiber-optic based biosensors have been reported useful for medical and environmental applications (Brogan and Walt, 2005; Bosch et al., 2007; Atias et al., 2009; Wolfbeis, 2006). Long period grating (LPG) fibers form a special class of transducers, and have been proposed for some biosensing applications (Kuhlmey et al., 2008; Barnes et al., 2008, 2010). The LPG fiber sensors produce highly sensitive resonance wavelength shift signals upon induced changes in the ambient refractive index to offer direct and label free detections. The ability of LPGs to couple light from the fiber core to the fiber cladding helps in direct probing of the surrounding media. Any change in the refractive index of the surrounding media results into spectral shifts, whose demodulation and proper correlations leads to highly sensitive quantitative chemical analyses. Apart from the already cited references, some other important citations on the applications of LPG fibers report the detection of antigen (DeLisa et al., 2000), pH (Goicoechea et al., 2008) and medically relevant parameters (Mishra et al., 2011). Immunosensing with LPG requires their surface modification with molecular recognition elements, such as antibodies, enzyme, etc. Main methods for the immobilization of bioreceptors on the LPG surface ∗ Corresponding author. Tel.: +91 172 2659951; fax: +91 172 2659951. E-mail address: [email protected] (U. Tiwari). 0956-5663/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2011.12.051 include adsorption (Liu et al., 2000), ionic bonding by electrostatic self-assembly technique (Elosúa et al., 2006; Wang et al., 2009), cross-linking by means of multifunctional reagent (Lee and Walt, 2000), covalent bonding (Disley et al., 1998; Stanford et al., 2009), and avidin–biotin interaction (Marks et al., 2002). Covalent binding is considered as the most effective of all the above quoted attachment strategies. This approach allows the active sites to remain unobstructed and chemically reactive. Some of the extremely useful inherent properties of the LPG fibers, such as their high sensitivity and smart transduction instrumentation, have motivated us to exploit them for enzyme based immunosensing of glucose. Presently, only electrochemical methods are more popular for the glucose analysis. Proposed optical based biosensing of glucose with LPG fibers may contribute to the much required developments of alternative detection devices, which can also compete with generally trusted laboratory tests. The work described herein first time demonstrates the feasibility of enzyme coated LPG fibers for the photonics based glucose sensing. Since the successful and precise working of the enzyme coated LPG sensor depends entirely upon the quality of the fiber’s coverage with the protein molecule, the detailed spectroscopic and microscopic investigations have been undertaken to confirm the effectiveness of the adopted covalent immobilization methodology. In some published reports, unmodified LPG fibers have been proposed for the non-selective analysis of very highly concentrated glucose solutions, e.g. 300 mg mL−1 (Libish et al., 2011; Kim et al., A. Deep et al. / Biosensors and Bioelectronics 33 (2012) 190–195 2005). Dilute analyte solution tendered insignificant changes in the refractive index and no wavelength shifts could be recorded to deliver usable data. In our approach, the immobilization of a specific enzyme has caused the conversion of glucose to gluconic acid during the analysis. This reaction props up change in the media’s refractive index even with dilute (e.g. 0.1–3 mg mL−1 ) glucose solutions, which, in turn, changes the position of the transmission peak in the associated optical spectrum. Wavelength shifts are agreeably correlated within certain range of analyte concentration. 2. Materials and methods Analytical grade reagents and sterile deionized water were used for preparing all the working solutions. d-glucose, glucose oxidase (GOD, biological source: fungus Aspergillus niger), peroxidase (POD), o-dianisidine and 3-aminopropyl-triethoxysilane (APTES) were purchased from Sigma–Aldrich, India. Long Period Grating fibers (periodicity 550 m) were fabricated by exposing hydrogenated standard single mode fiber core to KrF laser using point-by-point method. Prior to further use, the LPG fibers were cleaned by immersion in 100 mL of 5% nitric acid for 2 h at 90 ◦ C followed by thorough rinsing with deionized water and ethanol. The hydroxyl-groups of the cleaned fibers were activated by exposing them with 95% (v/v) sulphuric acid for 1 h at room temperature. After drying in a convection oven for 4 h at 115 ◦ C, the LPG fibers were left in contact with 10% (v/v) ethanolic solution of aminopropyl triethoxysilane (APTES) for 30 min at room temperature. It was followed by rinsing with water and ethanol to remove non-covalently adsorbed silane compounds. The above reaction produced free amine groups on the fiber surface. All the used buffers (acetic acid/sodium acetate and sodium dihydrogen phosphate/disodium hydrogen phosphate) were prepared in compositions as calculated with the help of the software ‘Buffer Maker Version 1.0.1.55’. The ionic strengths of different buffers were forced to a constant value of 50 mM. Immobilization of GOD on the NH2 functionalized optical fibers was achieved by immersing the fibers in 5 mg ml−1 buffered solution (sodium acetate) of glucose oxidase (GOD) for 100 min. Enzyme immobilized fibers were washed with sodium acetate buffer (SA) and then dried in the air. To investigate any change in the GOD activity after immobilization on the optical fibers, a reaction mixture (S1) of 0.1 mL glucose solution (5 mM), 20 L peroxidase and 0.05 mL of o-dianisidine (1%, w/w) was prepared in 3 mL of phosphate buffer medium (50 mM, pH 7.5). Immobilized fibers were cut in to number of pieces and then their weighed (15 mg–40 mg) amounts were added into 500 L of the prepared reaction mixture‘S1’. These contents (in micro centrifuge vials) were gently mixed in a temperature controlled (35 ± 1 ◦ C) shaker for 1 h. After the reaction, 50 L of the liquid was pipetted out and tested for absorbance (at 500 nm) on a spectrophotometer (UV–Vis–NIR, Varian, Cary 5000). Consistency in the coating of APTES and GOD on the optical fibers was investigated by Field Emission Scanning Electron Microscope (FESEM, Hitachi S4300 SE/N) and Confocal Laser Scanning Microscope (CLSM, Zeiss, LSM 510). Chemical modification of the fibers with NH2 functional groups was characterized with Infrared (FTIR, Nicolet iS10) and Raman (Invia Raman, Renishaw) spectrometers. For glucose analysis, the enzyme immobilized LPG sensor was fixed between the two micropositioner stages to maintain the conditions of constant temperature, pressure and strain throughout the experiment. One end of the fiber was connected to white light source while the other was connected to an optical spectrum analyzer (OSA, Yokogawa AQ6319, Ando electric Co. Ltd., Japan) for constant monitoring of the transmission spectrum. A vertical 191 moving stage was used to dip the LPG in different test solutions. The sensor was exposed to the different analyte concentrations (0.1, 0.5, 1.0, 1.5, 2.0, 3.0, 4.0, 5.0 mg ml−1 glucose in sodium acetate buffer of pH 5.6) and the resulting transmission spectra were collected. Simple refractive index measurements were done on Bellingham Stanley (RFM 840) system. 3. Results and discussion 3.1. Optimization of the fiber silanization and enzyme immobilization steps The silanization of LPG surface is a crucial step to ensure the efficiency and reproducibility of the subsequent enzyme immobilization step. Some preliminary experiments were carried out to work out best silanization conditions and to avoid excess aggregation. A fix concentration (10%) of the ethanolic APTES solution was used to treat the fibers for different durations. After every experiment, the refractive index of the residual APTES solution was checked and its value was associated with the unreacted amount of the reagent. The values of refractive index of residual APTES solution became constant (1.322 ± 0.001) after 25 min of the incubation of fibers in the reagent solution. For all further studies, the silanization of the fibers was carried out by keeping the incubation time to 30 min. Similarly, the time required for the immobilization of GOD on the silanized LPG was optimized by immersing the functionalized fibers in 5 mg ml−1 solution of GOD in sodium acetate buffer for different time durations. The medium was intermittently disturbed and manually stirred to maintain homogenous enzyme dispersion. Absorption data of the supernatant liquor collected after each set of the experiment (Fig. S1) revealed that the desired enzyme immobilization on the silanized fibers may need approximately 75 min of their incubation with GOD solution. For all further studies, the enzyme immobilization on the silanized LPG fibers was achieved by 100 min of incubation. Some experiments were carried out to test the conservation of enzyme activity after immobilization onto the optical fibers. In general, the glucose oxidase catalyses the oxidation of -d-glucose to d-glucono-␦-lactone with the concurrent release of hydrogen peroxide (H2 O2 ) (Eq. (1)). During the enzyme activity test, the presence of peroxidase (POD) causes the evolved H2 O2 to enter into a second reaction involving o-dianisidine (reduced) with the quantitative formation of a o-dianisidine (oxidized) dye complex (Eq. (2)) which is measured at 500 nm. GOD D-Glucose + O2 + H2 O−→D-Glucono-1,5-lactone + H2 O2 POD H2 O2 + o-Dianisidine (reduced)−→o-Dianisidine (oxidized) (1) (2) Fig. S1 (inset) also gives the absorption data for enzyme assay test carried out on the GOD immobilized fibers. Linearity of the reaction curve for various concentrations of glucose oxidase in the assay mixture is an evidence of the conservation of GOD activity after its immobilization on the optical fibers. 3.2. Surface characteristics of the silanized and enzyme immobilized fibers Fig. 1 shows the FESEM micrographs of the original, silanized and GOD immobilized optical fibers. Surface modification is evident. It is also clear that the protocols used for the silanization and the subsequent enzyme immobilization produce uniform and nicely adhered biomolecule layers. Some charging was observed during the scanning of GOD coated fiber, which may be due to the interaction of electron beam with some of the non-conducting organic groups present in the enzyme structure. 192 A. Deep et al. / Biosensors and Bioelectronics 33 (2012) 190–195 Fig. 1. (a) FESEM micrograph of the optical fiber, (b) FESEM micrograph of the silanized optical fiber, (c) FESEM micrograph of the enzyme immobilized optical fiber and (d) CLSM images of the GOD immobilized fluorescent optical fibers. Intrinsic fluorescent property of the GOD was utilized to optically scan the enzyme immobilized fibers under a confocal laser scanning microscope. A laser wavelength of 458 nm with (10% transmission) was used for the excitation. Laser scanning at 200× optical magnification yielded fluorescent images (Fig. 2d) of the optical fibers, giving fair evidence about the consistent enzyme immobilization. As a control experiment, APTES treated fibers were also monitored under the same conditions. No fluorescent signal was identifiable. Hence, the observed fluorescence in the fibers can be attributed to the presence of GOD molecules. molecules. Thus available NH2 groups bind with the enzyme’s COOH group through covalent interaction to yield the protein immobilization according to Eq. (3). 3.3. FTIR and Raman spectroscopic investigations FTIR and Raman data provide complementary information on various vibration transitions and vibrational bands to verify the efficiency of fiber functionalization and subsequent GOD grafting processes. Fig. 2 shows the FTIR spectra of the silanized (solid lines) and GOD immobilized (dotted lines) optical fibers. The appearance of peaks at 1538 and 1634 cm−1 signifies the vibration of primary amine NH2 and N H scissoring bending. This is further supported by strong stretching at 3428 cm−1 . Attachment of aliphatic groups is highlighted by C H asymmetric and symmetric stretching modes at 2854 and 2925 cm−1 . Furthermore, absence of free Si OH functional groups (no peaks at 795 or 1177 cm−1 ) indicate that APTES molecules almost smoothly cover the fiber surface and provide NH2 functional groups to support further immobilization of GOD. The attachment of enzyme is characterized with the appearance of a peak at 1637 cm−1 depicting the N C vibrations. A weak band at 1559 cm−1 also suggests the interaction between the NH2 group of the APTES treated optical fiber and the carboxyl group of GOD. A broad and intense band at 3448 cm−1 confirms the N H stretching mode. The presence of two weak bands at 2825 and 2930 cm−1 is also suggesting the successful grafting of the protein Eq. (1) depicts the scheme of GOD immobilization on the fiber surface. Sulphuric acid treatment of the cleaned fibers activates OH groups that have further been used for the assembly of APTES Fig. 2. FTIR spectra of the silanized (solid line) and enzyme immobilized (dotted line) optical fibers. (3) A. Deep et al. / Biosensors and Bioelectronics 33 (2012) 190–195 Fig. 3. Raman spectral studies of (a) silanized optical fibers, (b) native GOD and (c) GOD immobilized optical fibers. components on to the optical fiber. The above FTIR studies prove that GOD molecules are well bonded on the silanized fiber surface. Raman spectra of the silanized and enzyme coated fibers were collected at a laser excitation wavelength of 785 nm. Fig. 3a shows the recorded data for the silanized fiber in two different ranges, 1200–1700 cm−1 and 2700–3000 cm−1 , for the sake of better peak presentations. Bands observed at 1639 and 1336 cm−1 are attributed to the NH2 group and NH with CH stretch modes, respectively. Bands at 1459 and 1414 cm−1 suggest the presence of CH2 scissoring mode and CN, respectively. It indicates the involvement of silane molecules belonging to APTES attached at the surface of the fiber. A distinct Raman band at 1314 cm−1 accounts for the C C stretch mode. The broad Raman bands at 2906, 2921 and 2967 cm−1 are ascribed to CH stretching modes due to the attachment of APTES molecules. Other weak and broad Raman bands at 3300–3500 cm−1 are assigned to the asymmetric and symmetric modes of NH2 stretching. Raman studies have verified that the APTES treatment to the optical fibers yields in the generation of free amine groups, which can further be utilized for the GOD attachment. Fig. 3b and c show the Raman spectra of the native GOD and the GOD immobilized optical fiber, respectively. Spectrum 3b reveals characteristic Raman bands for the native GOD molecule at 409, 632, 682, 863, 999, 1262, 1341, 1535 and 1775 cm−1 . Observed bands at 682, 409 and 632 cm−1 indicate the presence of S atom, ı (C S) deformations and (C S) stretching, respectively. CNC symmetric stretching of amine groups is highlighted by a band at 863 cm−1 . C O, C C stretching, p(NH2 ) rocking and NH2 twisting vibrations lead to the appearance of Raman bands at 1262 and 1535 cm−1 . Significant bands at 1535 and 1775 cm−1 may be assigned to the C C stretching and C O stretching vibrations, respectively. Raman spectrum of the enzyme immobilized fiber (Fig. 3c) reveals almost all the peaks of the native GOD molecule. This observation suggests that the main features of the native glucose oxidase are conserved after being immobilized 193 Fig. 4. Transmission spectra and detailed graphical illustration showing variation of wavelength shift as a function of varying glucose concentrations (0.1–5 mg mL−1 ). Relative standard deviation (RSD) for 3 different sample measurements: 0.015 < RSD < 0.030. on the fiber surface. Generally amide II band remain inactive or very weak in the Raman spectrum, however, notable emergence of a weak band at 1670 cm−1 (amide I) is suggestive of APTES-GOD covalent interaction. FTIR and Raman spectral studies have successfully revealed that the LPG fibers can be conveniently modified by APTES to generate free NH2 groups, which effectively function as enzyme binding sites. These modified optical fibers have further been coupled with measurement system to test their usefulness for glucose detection. 3.4. Glucose detection with GOD immobilized LPG fibers Immobilization of the enzyme on the LPG fibers has rendered a sensing platform, which not only provides specificity towards glucose detection but also protect the fiber from detecting minor refractive index changes in the presence of possible interfering species. Fig. S2 schematically represents the experimental setup with GOD immobilized LPG fiber (exposed length = 2 cm). A blank solution (2 mL) of sodium acetate buffer was first transferred in to the sample cell and the stable optical spectrum collected (Fig. S3). The observed peak position was referred as reference. The buffer was then carefully removed from the sample cell, and a glucose solution (in sodium acetate buffer) was introduced for analysis. The conversion of -d-glucose to d-glucono-␦-lactone (which is finally hydrolyzed to gluconic acid) caused the change in the refractive index of the medium, thus inducing wavelength shift in the optical spectrum. A stable signal (Fig. S3) was recordable after 30 s of the reaction. Subsequently, glucose solutions with different physiologically important concentrations (0.1–5 mg mL−1 ) were analyzed in the above manner. The obtained optical spectra are presented in Fig. 4. Data for only 3 representative concentrations have been plotted for the sake of clear graphical illustration, which shows 194 A. Deep et al. / Biosensors and Bioelectronics 33 (2012) 190–195 Table 1 Analysis of glucose from serum samples with glucose oxidase immobilized LPG fiber. Sample Glucose content as per hospital test (in mg mL−1 ) Wavelength shift on GOD-LPG sensor w.r.t. serum plasma sample having known 1.05 mg mL−1 glucose content (nm) Corresponding experimental glucose content (mg mL−1 ) Estimated glucose content after incorporating reference data (column 4 + 1.05 mg mL−1 ) % Variation w.r.t. hospital data S-1 S-2 S-3 1.52 1.97 2.72 1.2 ± 0.02 1.8 ± 0.02 2.2 ± 0.03 0.55 1.0 1.6 1.60 2.05 2.65 +5.26 +4.06 to 2.57 transmission peaks with wavelength shifts at two positions, near 1620 and 1550 nm. The signals obtained near 1620 nm provided the desired pattern of wavelength shifts with respect to the changing glucose concentration. Detailed information on the wavelength shifts (with respect to control transmission peak at 1620 nm) as a function of glucose concentrations is also plotted in Fig. 4. It is clear that the GOD immobilized LPG fiber give fairly linear response in the glucose concentration range of 0.1–3 mg mL−1 . The assay test with higher concentrations of glucose did not give proper response possibly due to the fact that the selected length (2 cm) of the LPG fiber did not have enough enzyme loading to support the complete oxidation of the test solution. However, it must be noted that the above sensor in the said conditions is suitable for the analysis of physiologically important glucose concentrations (10–300 mg dL−1 ). All the above described experiments were conducted at least three times with different sets of enzyme immobilized optical fibers, and the observed average variation in the wavelength shifts was limited to ±0.1 nm. The glucose detection with GOD immobilized LPG fibers has been proved to be very sensitive and the technique’s usability is demonstrated for the analysis of dilute (10–300 mg dL−1 ) glucose concentrations. Though it has been shown earlier that the unmodified LPG fibers may also be used for detecting glucose concentrations (Libish et al., 2011; Kim et al., 2005), those reported studies dealt with much higher (e.g. 30 g mL−1 ) levels of the analyte. Our investigations (Fig. S4) on a simple LPG fiber have shown insignificant optical response with respect to the analysis of dilute glucose solutions (0.1–5 mg mL−1 ). It can be correlated with minuscule changes in the medium’s refractive index when working with dilute solutions (Fig. S5). Glucose analysis on GOD modified LPG fiber is associated with the conversion of glucose to the gluconic acid; the evolution of which changes the medium’s refractive index. Study of solutions after the reaction indicated wider changes in the refractive index (Fig. S5) accounting for the appearance of significant transmission shifts even at low glucose concentrations. Solution pH is known to be one of the most important factors affecting the enzyme activity and stability, and hence the sensor’s performance. Effect of pH on the LPG sensor performance was determined in the range of 5.0–8.0 against a constant glucose concentration of 2.0 mg mL−1 . Acetic acid/sodium acetate buffer was used for lower pH value experiments while sodium dihydrogen phosphate/disodium hydrogen phosphate buffer was employed for making analyte solutions of higher pH values. Fig. 5 shows that the performance of the LPG sensor was better in the pH range of 5.0–7.0. Higher pH values inhibit the enzyme activity, and therefore the wavelength shifts were comparatively less resolved. Response time of the sensor was assessed by recording several optical spectra during the analysis of 1.0 mg mL−1 glucose solution. The data of the study are also given in Fig. 5. Stable optical spectrum was recorded within 30 s, thereby demonstrating that the enzyme–substrate reaction took little time to complete and yield the stable result. The different detailed investigations have identified the optimized experimental parameters for the analysis of glucose (10–300 mg dL−1 ) under certain test conditions (exposed length of LPG = 2 cm, GOD for immobilization = 5 mg mL−1 , phase volume = 2 mL). However, these parameters are selectable according to the specific requirements. In such a case, the calibration of the Fig. 5. Effects of pH and reaction time on the optical signal during the analysis of 2.0 mg mL−1 glucose on GOD immobilized optical fibers. Relative standard deviation (RSD) for 3 different sample measurements: 0.020 < RSD < 0.035. system will need to be redrawn. The practical utility of the proposed LPG sensor was evaluated by analyzing the glucose content in serum samples. Fresh and clinically analyzed plasma samples were collected from the local hospital. These were diluted before introducing into the LPG setup. A plasma sample with known (1.05 mg ml−1 ) glucose content was first analyzed to record a reference optical spectrum. Subsequent analyses of the samples of the diabetic patients were analyzed and the results are summarized in Table 1. The glucose content of the blood was calculated from the calibration curve. The results were satisfactory and agreed closely with those measured by the biochemical analyzer done by local hospital. 4. Conclusions The presented research data demonstrate the successful immobilization of GOD on to the silanized LPG fibers for the development of a new glucose sensing technique. The enzyme immobilization has been achieved by modifying the fiber surface with APTES. Free amine groups of the APTES bind with carboxyl groups of GOD to render a sensing probe. Microscopic studies prove the uniform and effective coating of the optical fiber with the protein. FTIR and Raman spectroscopic studies have further revealed the efficacy of the undertaken strategy to develop a LPG based sensor by giving detailed information on various vibrations transitions and bands after each step. FTIR, Raman spectroscopy and Confocal microscopy investigations provide new dimensions to the LPG based sensor fabrication, which may also be usefully referred for other similar purposes. During the immobilization process, intermittent washing and rinsing steps ensured the removal of extra materials attached through non-specific interactions. The contact of different glucose solutions with the enzyme immobilized LPG fiber leads to patterned wavelength shifts. A practical range of signal linearity implies the usefulness of the proposed sensor for the detection of physiologically significant glucose concentrations. A. Deep et al. / Biosensors and Bioelectronics 33 (2012) 190–195 Development of simple, inexpensive, direct and real-time monitoring glucose sensors is of continuous interest as the traditional methods for glucose concentration measurement require different laboratorial analyses. LPG fiber based sensitive photonic biosensing may offer an attractive solution in this respect due to some intrinsic merits, such as high sensitivity, immunity to electromagnetic interference, small size, fast response etc. Some other important advantages of the proposed optical technique over other label free amperometric, potentiometric and conductometric metal/dielectric platforms include the use of much efficient and convenient signal transduction protocol and low signal to noise ratio. The proposed technique of protein immobilization could also be useful to develop optical fiber based sensors for applications in various other fields like disease diagnosis, pathogen testing, study of environmental pollutants, etc. Acknowledgement Thanks are due to the Council of Scientific and Industrial Research, India (NWP 035 & NWP 026) for the financial support to the work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2011.12.051. 195 References Atias, D., Liebes, Y., Chalifa-Caspi, V., Bremand, L., Lobel, L., Marks, R.S., Dussart, P., 2009. Sens. Actuators B, 206–215. Barnes, J., Dreher, M., Plett, K., Brown, R.S., Crudden, C.M., Loock, H.-P., 2008. Analyst 133 (11), 1541–1549. Barnes, J.A., Brown, R.S., Cheung, A.H., Dreher, M.A., Mackey, G., Loock, H.P., 2010. Sens. Actuators B 148 (1), 221–226. Bosch, M., Sánchez, A., Rojas, F., Ojeda, C., 2007. Sensors 7 (6), 797–859. Brogan, K.L., Walt, D.R., 2005. Curr. Opin. Chem. Biol. 9 (5), 494–500. DeLisa, M.P., Zhang, Z., Shiloach, M., Pilevar, S., Davis, C.C., Sirkis, J.S., Bentley, W.E., 2000. Anal. Chem. 72 (13), 2895–2900. Disley, D.M., Blyth, J., Cullen, D.C., You, H.-X., Eapen, S., Lowe, C.R., 1998. Biosens. Bioelectron. 13 (3–4), 383–396. Elosúa, C., Bariáin, C., Matías, I.R., Arregui, F.J., Luquin, A., Laguna, M., 2006. Sens. Actuators B 115 (1), 444–449. Goicoechea, J., Zamarreño, C.R., Matías, I.R., Arregui, F.J., 2008. Sens. Actuators B 132 (1), 305–311. Kim, D.-W., Zhang, Y., Cooper, K.L., Wang, A., 2005. Appl. Opt. 44 (26), 5368–5373. Kuhlmey, B.T., Luan, F., Lazaro, J.M., Fu, L., Eggleton, B.J., Yeom, D.-I., Coen, S., Wang, A., Knight, J.C., 2008. AIP Conf. Proc. 1055 (1), 61–64. Lee, M., Walt, D.R., 2000. Anal. Biochem. 282 (1), 142–146. Libish, T.M., Linesh, J., Bobby, M.C., Nithyaja, B., Mathew, S., Pradeep, C., Radhakrishnan, P., 2011. Sens. Transducers 129 (6), 142–148. Liu, X., Farmerie, W., Schuster, S., Tan, W., 2000. Anal. Biochem. 283 (1), 56–63. Marks, R., Novoa, A., Thomassey, D., Cosnier, S., 2002. Anal. Biochem. 374 (6), 1056–1063. Mishra, V., Singh, N., Tiwari, U., Kapur, P., 2011. Sens. Actuators A 167 (2), 279–290. Stanford, C.J., Ryu, G., Dagenais, M., Hurley, M.T., Gaskell, K.J., DeShong, P., 2009. J. Sens. Wang, Z., Heflin, J.R., Van Cott, K., Stolen, R.H., Ramachandran, S., Ghalmi, S., 2009. Sens. Actuators B 139 (2), 618–623. Wolfbeis, O.S., 2006. Anal. Chem. 78 (12), 3859–3874.