Biosensors and Bioelectronics 59 (2014) 293–299
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Highly sensitive and selective glucose sensor based on ultraviolet-treated nematic liquid crystals Shenghong Zhong, Chang-Hyun Jang n Department of Chemistry, Gachon University, Seongnam-Si, Gyeonggi-Do 461-701, Korea
art ic l e i nf o
a b s t r a c t
Article history: Received 11 January 2014 Received in revised form 18 March 2014 Accepted 31 March 2014 Available online 8 April 2014
Glucose is an extremely important biomolecule, and the ability to sense it has played a significant role in facilitating the understanding of many biological processes. Here, we report a novel glucose sensor based on ultraviolet (UV)-treated nematic liquid crystals. Submerging UV-treated 4-cyano-40 -pentylbiphenyl (5CB) in a glucose solution (while carefully adjusting its pH to 7.5 with NaOH and HCl) triggered an optical response, from dark to bright, observed with a polarized microscope. Notably, 5CB was located inside a glucose oxidase (GOx)-modified gold grid. We exploited this pH-driven phenomenon to design a new glucose sensor. This device could detect as little as 1 pM analyte, which is 3 orders of magnitude lower than the detection limit of the most sensitive glucose sensor currently available. It also exhibits high selectivity due to GOx modification. Thus, this is a promising technique for glucose detection, not only for clinical diagnostics, but also for sensing low levels of glucose in a biological environment (e.g., single cells and bacterial cultures). & 2014 Elsevier B.V. All rights reserved.
Keywords: Biosensors Glucose oxidase Liquid crystals PH-driven phenomenon Optical response
1. Introduction Glucose-monitoring technology has been used in the management of diabetes for more than three decades. Recently, interest in glucose sensors has been stimulated by their possible applications in the food industry (Terry et al., 2005), and for biological, chemical, and pharmaceutical analyses (Bankar et al., 2009). Many technologies are being pursued to develop high-performance glucose sensors, including surface plasmon resonance (SPR) (Aslan et al., 2004), carbon nanotubes (CNTs) (Rong et al., 2007), electrochemical signal transduction (EST) (Chen et al., 2000, 2006; Sapre et al., 2000), organic electrochemical transistors (OECT) (Liao et al., 2013), fluorescence signal transmission (FST) (He et al., 2006; Meadows and Schultz, 1993; Tang et al., 2008), and ZnO nanorod arrays (ZNA) (Liu et al., 2009). Some of these sensors exhibit very impressive sensitivities (e.g., 1 nM). Nevertheless, they are still not sensitive enough to detect trace quantities of glucose resulting from some biological processes, such as protein biosynthesis or cellular signal transduction. The electrically induced reorientation properties and optical orientation response of liquid crystals (LCs) have held the attention of many researchers. Some novel applications, such as LC lasers (Coles and Morris, 2010), organic light-emitting diodes (OLEDs) (Aldred et al., 2005), and biosensors (Hu and Jang, 2012, 2011; Lee et al., 2013; Lin et al., 2011; Liu et al., 2012), have been reported. n
Corresponding author. Tel.: þ 82 31 750 8555. E-mail address: [email protected] (C.-H. Jang).
http://dx.doi.org/10.1016/j.bios.2014.03.070 0956-5663/& 2014 Elsevier B.V. All rights reserved.
In order to improve the detection limit of glucose, intensive efforts have been made in the exploration of sensors based on LCs, including liquid-crystal polarization modulators (Lo and Yu, 2006) and LC membranes (Rowinski et al., 2008). LC membranes can also be used to determine the absolute configuration of glucose (James et al., 1996, 1993). In these studies, James et al. demonstrated that when D-glucose bound to the cholesteryl boronic acid moieties that were composited in LCs, the color of the system turned red, whereas L-glucose turned the LC color blue. However, the long time required for glucose to diffuse into the LC layers has limited the applicability of this system. Here, we report a new glucose sensor, which provides rapid optical responses to the orientational transition of LCs occurring at the aqueous/LC interface. Inspired by the pH-dependence phenomenon of carboxylic acid-doped 5CB (Kinsinger et al., 2007; Park et al., 2006) and the high sensitivity for monitoring enzymatic reactions of LCs (Bi et al., 2009), we developed a novel LC-based glucose sensor that displays orientation transitions with as little as 1 pM glucose. Furthermore, this glucose sensor can be employed to measure glucose concentrations between 1 pM and 50 nM.
2. Experimental 2.1. Materials Nematic liquid crystal 4-cyano-40 -pentylbiphenyl (5CB), manufactured by BDH, was purchased from EM industries (Hawthorne,
294
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
NY). The premium glass microscope slides were obtained from Fisher Scientific (Pittsburgh, PA). Gold specimen grids were purchased from GILDER. Hydrochloric acid solution (0.1 M) was purchased from FLUKA. n-Heptane (anhydrous) was purchased from Daejung Chemicals & Metals Co. Ltd. (South Korea). Sulfuric acid, hydrogen peroxide (30% w/v), sodium hydroxide, phosphatebuffered saline (PBS) (10 mM phosphate, 138 mM NaCl, 2.7 mM KCl; pH 7.4), octyltrichlorosilane (OTS), L-ascorbic acid, 11mercaptoundecanoic acid (MUA), N-hydroxysuccinimide (NHS), 1-ethyl-2-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC), D-( þ)-glucose, glucose oxidase (GOx), sucrose, D-( )-lactic acid, L-( þ)-lactic acid, and sodium dodecyl sulfate (SDS), were purchased from Sigma-Aldrich (St. Louis, MI). All aqueous solutions were prepared with deionized (DI) water (18.2 MΩ cm 1), using a Milli-Q water purification system (Millipore; Bedford, MA). 2.2. Treatment of glass microscope slides with OTS The glass microscope slides were cleaned using “piranha solution” (70% H2SO4 /30% H2O2) for 30 min at 80 1C under a stream of gaseous N2 (0.2 MPa). (Caution: “piranha solution” reacts violently with organic materials and should be handled with extreme caution; do not store the solution in a closed container.) The slides were then rinsed with deionized water, ethanol, and methanol (5 times each), and dried under a stream of gaseous N2, followed by heating to 120 1C overnight prior to OTS deposition. The piranha-cleaned glass slides were immersed in an OTS (0.12 mL)/n-heptane (60 mL) solution for 30 min at room temperature. The slides were then rinsed with methylene chloride (5 times) and dried under a stream of N2. 2.3. Preparation of GOx-modified gold grids We prepared the GOx-modified gold grids using the classical NHS/EDC method (Lahiri et al., 1999). Briefly, gold grids were cleaned using “piranha solution”. Self-assembled monolayers (SAMs) were formed by immersion of the gold grids in 2 mM ethanolic solution of MUA for 2 h, and then rinsed with a copious amount of ethanol and DI water. An aqueous solution of NHS/EDC (50 mM/200 mM) was used to activate the carboxylic acidterminated SAMs formed on the gold surface. After 30 min of reaction, the gold grids were rinsed with DI water and immersed into a GOx solution (2 mg/mL) for 6 h at 4 1C. Next, the GOxmodified grids were rinsed with DI water and dried under nitrogen gas at room temperature. The modified grids were used to anchor the liquid crystals. 2.4. Photochemical degradation of 5CB using UV light We prepared UV-treated 5CB using a method similar to a previously reported procedure (Park et al., 2006). Briefly, 5CB was placed in an 8-mL beaker under a Spectroline EN-280L longwave ultraviolet lamp (365 nm) equipped with two 8-W tubes and a filter assembly. The distance between the 5CB and the lamp was approximately 8 cm, and the entire system was placed in a box to block external light. Samples were prepared by exposing 0.6 g 5CB to the UV light for 48 h. 2.5. Preparation of optical cells The OTS-coated glass slides were fixed at the bottom of an eight-well chamber. TEM specimen grids were then placed onto the OTS-coated glass slide. Approximately 2.5 μL UV-tailored 5CB (isotropic state) was dispensed onto each grid, and the excess LC was removed by contacting a 20-μL capillary tube to the 5CB
droplet on the grid. Subsequently, the optical cell was immersed in 500 μL aqueous solution of interest at room temperature. 2.6. Glucose oxidase enzymatic reaction After immersion in 500 μL glucose solution (while carefully adjusting its pH to 7.5 with NaOH and HCl), the optical cell was placed into an oven maintained at 40 1C. The incubation time for higher concentrations (higher than 0.1 mM) was 20 min, and that for lower concentrations (lower than 50 nM) was 2 h. Then, the optical cells were slowly cooled to room temperature. 2.7. Optical examination of LC textures A polarized light microscope (ECLIPSE LV100POL; Nikon; Tokyo, Japan) was used to capture images of the optical textures observed by polarized light transmitted through the optical cells filled with UV-treated 5CB. All images were obtained using a 4 objective lens between a crossed polarizer and analyzer. Imaging the LC was accomplished with a digital camera (DS-2Mv; Nikon; Tokyo, Japan) attached to the polarized light microscope. The images were captured at a resolution of 1600 1200 pixels, a gain of 1.00 , and a shutter speed of 1/10 s.
3. Results and discussion 3.1. Sensing mechanism β-D-glucose can be oxidized to D-glucono-δ-lactone, which will then be non-enzymatically hydrolyzed to gluconic acid and H2O2 by glucose oxidase (GOx) (Witt et al., 2000). Most electrochemical glucose sensors rely on H2O2 to monitor glucose; however, we considered that it would be advantageous to use both H2O2 and gluconic acid, as this should theoretically improve sensitivity. Since both molecules are acidic, the enzymatic reaction would be expected to increase the acidity of the solution. In addition, carboxylic acid-doped LCs yield a rapid response to small pH changes (Bi et al., 2009). In general, at a higher pH value, a dark image will be observed, while a lower pH will result in a bright texture. Moreover, when 5CB is exposed to UV light, a photochemical reaction occurs to produce 4-cyano-40 -biphenylcarboxylic acid (CBA). These suggest that UV-treated 5CB should exhibit pH-dependent optical signals. When UV-treated 5CB confined to a gold grid was immersed in a phosphate-buffered saline (PBS) solution (pH 6.0), a bright image with good fan-shape was observed under a cross-polarized microscope (Fig. 1A). We still observed a bright image even after we increased the pH value to 7.4 (Fig. 1B). Interestingly, a slight increase to pH 7.5 immediately resulted in a dark image (Fig. 1C). After increasing the pH to 9.0, the image remained dark (Fig. 1D). In conclusion, a nominal change of pH from 7.4 to 7.5 triggered a bright-to-dark image transition. It is well known that planar orientations of LCs at the aqueous/ LC interface show bright images under a polarized microscope, whereas dark images are caused by homeotropic orientations. Moreover, previous studies have shown that when the adsorption of surfactants at the aqueous/LC interface exceeds a critical value, a planar-to-homeotropic orientational transition occurs (Bi et al., 2009). In our case, the density of CBA and 4-cyano-40 -biphenylcarboxylate (CB) at the aqueous/LC interface correlated with the pH value due to the ionic equilibrium of CBA. Increasing the pH will shift the equilibrium to the right and increase the density of CB but decrease the density of CBA. By contrast, decreasing the pH will shift the equilibrium to the left and decrease the density of CB, while increasing the density of CBA. Due to the amphiphilic nature
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
295
Fig. 1. Polarized microscopy images of UV-treated 5CB confined to gold grids were subjected to different pH PBS solutions: (A) pH ¼ 6.0, (B) pH¼ 7.4, (C) pH¼ 7.5, and (D) pH ¼9.0.
Fig. 2. (a) The schematic setup of the reported glucose sensor; (b) the sensing mechanism. With an unmodified gold grid, the pH of the aqueous solution was maintained at 7.5, sufficient surfactants cause a homeotropic orientation of UV-treated 5CB, resulting in a dark image. With the GOx-modified gold grid, the H þ released from the oxidation of β-D-glucose by GOx decreased the pH of the aqueous solution, a bright image results from carboxylate protonation and a homeotropic-to-planar orientation transition triggered by insufficient surfactants.
of CB and the hydrophobic nature of CBA (40 -Cyano-4biphenylcarboxylic acid | C14H9NO2 | ChemSpider, n.d.), increasing the density of CB (i.e., higher pH) will help introduce a homeotropic orientation and yield a dark image. By contrast, decreasing its density (i.e., lower pH) will facilitate a planar orientation and yield a bright image. As previously alluded to, the critical orientational transition corresponds to a specific pH value between 7.4 and 7.5 (Fig. 2b).
3.2. Monitoring glucose using an enzyme-modified gold grid The GOx-catalyzed oxidation of β-D-glucose is known to release H þ , thereby causing a slight increase in acidity that could be detected by UV-treated 5CB. Thus, we predicted that UV-treated
5CB could be used to monitor the presence of glucose. We constructed the detection system, as shown in the schematic of Fig. 2a, to verify our hypothesis. We first immobilized GOx on a gold grid via NHS/EDC conjugation chemistry (Lahiri et al., 1999). Then, we filled the meshes of the gold grid with UV-treated 5CB. Finally, the whole system was immersed in glucose solution (while carefully adjusting its pH to 7.5 with NaOH and HCl). After the enzymatic reaction was incubated at 40 1C for 2 h, the system was observed with a polarized microscope. The monitoring mechanism is illustrated in Fig. 2b. The H þ released from the oxidation of β-Dglucose by GOx triggered an optical response of the LCs from dark to bright. Fig. 3a-A shows that that the image of the LCs became bright in the presence of 10 pM glucose (pH 7.5). For comparison, the same system was immersed in pure PBS solution (pH 7.5) without glucose. Fig. 3a-B shows that the image remained dark.
296
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
Fig. 3. (a) Polarized light microscopy images of UV-treated 5CB were submerged in (A) glucose solution (pH 7.5) with a GOx-modified gold grid, (B) PBS solution (pH 7.5) with a GOx-modified gold grid, and (C) glucose solution (pH 7.5) with an unmodified gold grid. (b) Images of UV-treated 5CB were submerged in different clinical concentrations of glucose solutions (pH 7.5) after 20 min incubation at 40 1C, (A) 40 mM, (B) 10 mM, (C) 1 mM, and (D) 0.1 mM.
Fig. 4. Polarized light microscopy images of UV-treated 5CB submerged in a series of glucose solutions (pH 7.5) of varying concentrations: (A and G) 50 nM, (B and H) 1 nM, (C and I) 0.1 nM, (D and J) 10 pM, (E and K) 1 pM, and (F and L) 0.1 pM. The gold grids of A–F were modified by GOx, while those of G–L were unmodified.
Furthermore, we used an unmodified gold grid instead of the GOxmodified grid and immersed it in the same glucose solution; Fig. 3a-C shows the dark image obtained. In order to demonstrate the potential of this sensor for fast glucose detection in clinical diagnostics, we prepared a series of glucose solutions (pH 7.5) for which the concentration ranged from 0.1 mM to 40 mM. After 20 min incubation at 40 1C, bright images were obtained under a polarized microscope, as shown in Fig. 3b. These results suggest that this novel system can be used to monitor glucose owing to the H þ released during the enzymatic reaction.
3.3. Detection limit After confirming that glucose could be detected by UV-treated 5CB, we shifted our interest to the detection limit. We prepared a series of glucose solutions (pH 7.5), each with a different concentration. Subsequently, the UV-treated 5CB with a GOx-modified gold grid was immersed in each of these solutions. For comparison, we immersed the UV-treated 5CB with unmodified gold grids into these glucose solutions at the same time (Fig. 4). The images at the top (A–F) of Fig. 4 show the optical responses of UV-treated
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
297
Fig. 5. (a) Histograms of polarized light microscopy images of UV-treated 5CB with GOx-modified gold grids that were in contact with a series of glucose solutions (pH 7.5) of various concentrations. (b) Brightness value as a function of glucose concentration.
298
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
Fig. 6. Polarized microscopy images of UV-treated 5CB confined in the GOx-modified gold grids when placed in contact with different solutions (pH 7.5, 1 mM): (A) sucrose, (B) D-( )-lactic acid, (C) L-(þ )-lactic acid, (D) SDS, (E) L-ascorbic acid, and (F) glucose.
5CB confined in GOx-modified gold grids immersed in glucose solution, and the images at the bottom (G–L) show the responses of unmodified grids immersed in glucose solution. The following concentrations were used: 50 nM (images A and G), 1 nM (B and H), 0.1 nM (C and I), 10 pM (D and J), 1 pM (E and K), and 0.1 pM (F and L). The figure clearly indicates that the concentrations used in A–E produced bright images with good fan shape, while the others produced dark images without fan shape. This result suggests that the detection limit of this system is 1 pM. 3.4. Quantitative measurement of the concentration of glucose solution After identifying the detection limit of our system, we devoted our resources to developing an easier method to measure even lower glucose concentrations. It seems logical that lower glucose concentrations would yield darker images: the less H þ released during the enzymatic reaction means more surfactant (CB) at the aqueous/LC interface, resulting in darker images. Therefore, the analysis of image brightness is a potential method for measuring low glucose concentrations. We plotted a graph of the mean value (in color the term value means brightness) of the images against the logarithm of glucose solution concentration ranged from 1 pM to 50 nM. As shown in Fig. 5b, lower concentrations corresponded to lower brightness values (all the color information was analyzed using GIMP; see Fig. 5a for more details). The brightness of the images could be affected by many factors, including the thickness of the LCs, the depth and pH of the aqueous solution, the intensity of the incident light, and the structure of gold grids. Fortunately, industrial programmed processes can be employed to circumvent (or at least mitigate) these potentially problematic issues. Thus, this method could theoretically be used for measuring low concentrations of glucose with an acceptable degree of accuracy. Herein, we minimized the variation by fixing the illumination light intensity knob with tape, keeping the depth of aqueous solutions at 5 mm while adjusting the pH of all the solutions to 7.5. 3.5. Studying the selectivity of this novel sensor In addition to its extremely high sensitivity, this glucose sensor also displays high selectivity. We were able to exploit the fact that glucose is a type of saccharide, that it exhibits optical activity, that high acid content would affect the blood glucose sensing, and that the surfactant would affect the orientation of the LCs. For example, to demonstrate device selectivity, we examined the optical response of this glucose sensor when immersed in sucrose solution (Fig. 6A, a dark image), D-( )-lactic acid solution (Fig. 6B, a dark image), L-( þ)-lactic acid solution (Fig. 6C, a dark image), sodium dodecyl sulfate (SDS) solution (Fig. 6D, a dark image), and L-ascorbic acid solution (Fig. 6E, a dark image). The concentration of these solutions was 1 mM and the incubation time was 2 h. Fig. 6F shows that a bright image was produced when the sensor was immersed in glucose solution. These results suggest that our
sensor displays high selectivity due to the underlying enzymatic reaction it relies upon.
4. Conclusions In conclusion, we developed a novel glucose sensor based on UV-treated LCs. This sensor can be used not only for clinical glucose detection, but also to monitor low glucose concentrations. It shows outstanding performance with extremely high sensitivity and selectivity. The detection limit of 1 pM is 3 orders of magnitude better than the second most sensitive glucose sensor. These results hold great promise for developing a novel sensor for detecting trace quantities of glucose occurring in biological environments. Furthermore, these UV-treated liquid crystals could be used to monitor other molecules related to pH-changing bioreactions with high sensitivity. The ability to detect trace quantities of many biomolecules will greatly promote the development of several fields of research. Acknowledgment This study was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF2013R1A1A1A05008333) and a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (HI13C0891). References 40 -Cyano-4-biphenylcarboxylic acid | C14H9NO2 | ChemSpider, n.d. [WWW Document]. URL 〈http://www.chemspider.com/Chemical-Structure.2038713.html〉. Aldred, M.P., Contoret, A.E.A., Farrar, S.R., Kelly, S.M., Mathieson, D., O'Neill, M., Tsoi, W.C., Vlachos, P., 2005. Adv. Mater. 17, 1368–1372. Aslan, K., Lakowicz, J.R., Geddes, C.D., 2004. Anal. Biochem. 330, 145–155. Bankar, S.B., Bule, M.V., Singhal, R.S., Ananthanarayan, L., 2009. Biotechnol. Adv. 27, 489–501. Bi, X., Hartono, D., Yang, K.-L., 2009. Adv. Funct. Mater. 19, 3760–3765. Chen, T., Friedman, K.A., Lei, I., Heller, A., 2000. Anal. Chem. 72, 3757–3763. Chen, W., Yao, H., Tzang, C.H., Zhu, J., Yang, M., Lee, S.-T., 2006. Appl. Phys. Lett. 88, 213104. Coles, H., Morris, S., 2010. Nat. Photonics 4, 676–685. He, F., Tang, Y., Yu, M., Wang, S., Li, Y., Zhu, D., 2006. Adv. Funct. Mater. 16, 91–94. Hu, Q.-Z., Jang, C.-H., 2011. Colloids Surf. B Biointerfaces 88, 622–626. Hu, Q.-Z., Jang, C.-H., 2012. J. Biotechnol. 157, 223–227. James, T.D., Harada, T., Shinkai, S., 1993. J. Chem. Soc. Chem. Commun. 10, 857–860. James, T.D., Sandanayake, K.R.A., Shinkai, S., 1996. Angew. Chem. Int. Ed. Engl. 35, 1910–1922. Kinsinger, M.I., Sun, B., Abbott, N.L., Lynn, D.M., 2007. Adv. Mater. 19, 4208–4212. Lahiri, J., Isaacs, L., Tien, J., Whitesides, G.M., 1999. Anal. Chem. 71, 777–790. Lee, G., Carlton, R.J., Araoka, F., Abbott, N.L., Takezoe, H., 2013. Adv. Mater. 25, 245–249. Liao, C., Zhang, M., Niu, L., Zheng, Z., Yan, F., 2013. J. Mater. Chem. B 1, 3820–3829. Lin, I.-H., Miller, D.S., Bertics, P.J., Murphy, C.J., de Pablo, J.J., Abbott, N.L., 2011. Science 332, 1297–1300. Liu, H., Qian, X., Wang, S., Li, Y., Song, Y., Zhu, D., 2009. Nanoscale Res. Lett. 4, 1141–1145. Liu, Y., Cheng, D., Lin, I.-H., Abbott, N.L., Jiang, H., 2012. Lab Chip 12, 3746–3753. Lo, Y.-L., Yu, T.-C., 2006. Opt. Commun. 259, 40–48.
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
Meadows, D.L., Schultz, J.S., 1993. Anal. Chim. Acta 280, 21–30. Park, J.-S., Jang, C.-H., Tingey, M.L., Lowe, A.M., Abbott, N.L., 2006. J. Colloid Interface Sci. 304, 459–473. Rong, L.-Q., Yang, C., Qian, Q.-Y., Xia, X.-H., 2007. Talanta 72, 819–824. Rowinski, P., Rowinska, M., Heller, A., 2008. Anal. Chem. 80, 1746–1755.
299
Sapre, A.G., Bedekar, A., Deshpande, A.V., Lali, A.M., 2000. Biotechnol. Lett. 22, 569–573. Tang, B., Cao, L., Xu, K., Zhuo, L., Ge, J., Li, Q., Yu, L., 2008. Chemistry 14, 3637–3644. Terry, L.A., White, S.F., Tigwell, L.J., 2005. J. Agric. Food Chem. 53, 1309–1316. Witt, S., Wohlfahrt, G., Schomburg, D., Hecht, H.-J., Kalisz, H.M., 2000. Biochem. J. 347, 553–559.
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Highly sensitive and selective glucose sensor based on ultraviolet-treated nematic liquid crystals Shenghong Zhong, Chang-Hyun Jang n Department of Chemistry, Gachon University, Seongnam-Si, Gyeonggi-Do 461-701, Korea
art ic l e i nf o
a b s t r a c t
Article history: Received 11 January 2014 Received in revised form 18 March 2014 Accepted 31 March 2014 Available online 8 April 2014
Glucose is an extremely important biomolecule, and the ability to sense it has played a significant role in facilitating the understanding of many biological processes. Here, we report a novel glucose sensor based on ultraviolet (UV)-treated nematic liquid crystals. Submerging UV-treated 4-cyano-40 -pentylbiphenyl (5CB) in a glucose solution (while carefully adjusting its pH to 7.5 with NaOH and HCl) triggered an optical response, from dark to bright, observed with a polarized microscope. Notably, 5CB was located inside a glucose oxidase (GOx)-modified gold grid. We exploited this pH-driven phenomenon to design a new glucose sensor. This device could detect as little as 1 pM analyte, which is 3 orders of magnitude lower than the detection limit of the most sensitive glucose sensor currently available. It also exhibits high selectivity due to GOx modification. Thus, this is a promising technique for glucose detection, not only for clinical diagnostics, but also for sensing low levels of glucose in a biological environment (e.g., single cells and bacterial cultures). & 2014 Elsevier B.V. All rights reserved.
Keywords: Biosensors Glucose oxidase Liquid crystals PH-driven phenomenon Optical response
1. Introduction Glucose-monitoring technology has been used in the management of diabetes for more than three decades. Recently, interest in glucose sensors has been stimulated by their possible applications in the food industry (Terry et al., 2005), and for biological, chemical, and pharmaceutical analyses (Bankar et al., 2009). Many technologies are being pursued to develop high-performance glucose sensors, including surface plasmon resonance (SPR) (Aslan et al., 2004), carbon nanotubes (CNTs) (Rong et al., 2007), electrochemical signal transduction (EST) (Chen et al., 2000, 2006; Sapre et al., 2000), organic electrochemical transistors (OECT) (Liao et al., 2013), fluorescence signal transmission (FST) (He et al., 2006; Meadows and Schultz, 1993; Tang et al., 2008), and ZnO nanorod arrays (ZNA) (Liu et al., 2009). Some of these sensors exhibit very impressive sensitivities (e.g., 1 nM). Nevertheless, they are still not sensitive enough to detect trace quantities of glucose resulting from some biological processes, such as protein biosynthesis or cellular signal transduction. The electrically induced reorientation properties and optical orientation response of liquid crystals (LCs) have held the attention of many researchers. Some novel applications, such as LC lasers (Coles and Morris, 2010), organic light-emitting diodes (OLEDs) (Aldred et al., 2005), and biosensors (Hu and Jang, 2012, 2011; Lee et al., 2013; Lin et al., 2011; Liu et al., 2012), have been reported. n
Corresponding author. Tel.: þ 82 31 750 8555. E-mail address: [email protected] (C.-H. Jang).
http://dx.doi.org/10.1016/j.bios.2014.03.070 0956-5663/& 2014 Elsevier B.V. All rights reserved.
In order to improve the detection limit of glucose, intensive efforts have been made in the exploration of sensors based on LCs, including liquid-crystal polarization modulators (Lo and Yu, 2006) and LC membranes (Rowinski et al., 2008). LC membranes can also be used to determine the absolute configuration of glucose (James et al., 1996, 1993). In these studies, James et al. demonstrated that when D-glucose bound to the cholesteryl boronic acid moieties that were composited in LCs, the color of the system turned red, whereas L-glucose turned the LC color blue. However, the long time required for glucose to diffuse into the LC layers has limited the applicability of this system. Here, we report a new glucose sensor, which provides rapid optical responses to the orientational transition of LCs occurring at the aqueous/LC interface. Inspired by the pH-dependence phenomenon of carboxylic acid-doped 5CB (Kinsinger et al., 2007; Park et al., 2006) and the high sensitivity for monitoring enzymatic reactions of LCs (Bi et al., 2009), we developed a novel LC-based glucose sensor that displays orientation transitions with as little as 1 pM glucose. Furthermore, this glucose sensor can be employed to measure glucose concentrations between 1 pM and 50 nM.
2. Experimental 2.1. Materials Nematic liquid crystal 4-cyano-40 -pentylbiphenyl (5CB), manufactured by BDH, was purchased from EM industries (Hawthorne,
294
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
NY). The premium glass microscope slides were obtained from Fisher Scientific (Pittsburgh, PA). Gold specimen grids were purchased from GILDER. Hydrochloric acid solution (0.1 M) was purchased from FLUKA. n-Heptane (anhydrous) was purchased from Daejung Chemicals & Metals Co. Ltd. (South Korea). Sulfuric acid, hydrogen peroxide (30% w/v), sodium hydroxide, phosphatebuffered saline (PBS) (10 mM phosphate, 138 mM NaCl, 2.7 mM KCl; pH 7.4), octyltrichlorosilane (OTS), L-ascorbic acid, 11mercaptoundecanoic acid (MUA), N-hydroxysuccinimide (NHS), 1-ethyl-2-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC), D-( þ)-glucose, glucose oxidase (GOx), sucrose, D-( )-lactic acid, L-( þ)-lactic acid, and sodium dodecyl sulfate (SDS), were purchased from Sigma-Aldrich (St. Louis, MI). All aqueous solutions were prepared with deionized (DI) water (18.2 MΩ cm 1), using a Milli-Q water purification system (Millipore; Bedford, MA). 2.2. Treatment of glass microscope slides with OTS The glass microscope slides were cleaned using “piranha solution” (70% H2SO4 /30% H2O2) for 30 min at 80 1C under a stream of gaseous N2 (0.2 MPa). (Caution: “piranha solution” reacts violently with organic materials and should be handled with extreme caution; do not store the solution in a closed container.) The slides were then rinsed with deionized water, ethanol, and methanol (5 times each), and dried under a stream of gaseous N2, followed by heating to 120 1C overnight prior to OTS deposition. The piranha-cleaned glass slides were immersed in an OTS (0.12 mL)/n-heptane (60 mL) solution for 30 min at room temperature. The slides were then rinsed with methylene chloride (5 times) and dried under a stream of N2. 2.3. Preparation of GOx-modified gold grids We prepared the GOx-modified gold grids using the classical NHS/EDC method (Lahiri et al., 1999). Briefly, gold grids were cleaned using “piranha solution”. Self-assembled monolayers (SAMs) were formed by immersion of the gold grids in 2 mM ethanolic solution of MUA for 2 h, and then rinsed with a copious amount of ethanol and DI water. An aqueous solution of NHS/EDC (50 mM/200 mM) was used to activate the carboxylic acidterminated SAMs formed on the gold surface. After 30 min of reaction, the gold grids were rinsed with DI water and immersed into a GOx solution (2 mg/mL) for 6 h at 4 1C. Next, the GOxmodified grids were rinsed with DI water and dried under nitrogen gas at room temperature. The modified grids were used to anchor the liquid crystals. 2.4. Photochemical degradation of 5CB using UV light We prepared UV-treated 5CB using a method similar to a previously reported procedure (Park et al., 2006). Briefly, 5CB was placed in an 8-mL beaker under a Spectroline EN-280L longwave ultraviolet lamp (365 nm) equipped with two 8-W tubes and a filter assembly. The distance between the 5CB and the lamp was approximately 8 cm, and the entire system was placed in a box to block external light. Samples were prepared by exposing 0.6 g 5CB to the UV light for 48 h. 2.5. Preparation of optical cells The OTS-coated glass slides were fixed at the bottom of an eight-well chamber. TEM specimen grids were then placed onto the OTS-coated glass slide. Approximately 2.5 μL UV-tailored 5CB (isotropic state) was dispensed onto each grid, and the excess LC was removed by contacting a 20-μL capillary tube to the 5CB
droplet on the grid. Subsequently, the optical cell was immersed in 500 μL aqueous solution of interest at room temperature. 2.6. Glucose oxidase enzymatic reaction After immersion in 500 μL glucose solution (while carefully adjusting its pH to 7.5 with NaOH and HCl), the optical cell was placed into an oven maintained at 40 1C. The incubation time for higher concentrations (higher than 0.1 mM) was 20 min, and that for lower concentrations (lower than 50 nM) was 2 h. Then, the optical cells were slowly cooled to room temperature. 2.7. Optical examination of LC textures A polarized light microscope (ECLIPSE LV100POL; Nikon; Tokyo, Japan) was used to capture images of the optical textures observed by polarized light transmitted through the optical cells filled with UV-treated 5CB. All images were obtained using a 4 objective lens between a crossed polarizer and analyzer. Imaging the LC was accomplished with a digital camera (DS-2Mv; Nikon; Tokyo, Japan) attached to the polarized light microscope. The images were captured at a resolution of 1600 1200 pixels, a gain of 1.00 , and a shutter speed of 1/10 s.
3. Results and discussion 3.1. Sensing mechanism β-D-glucose can be oxidized to D-glucono-δ-lactone, which will then be non-enzymatically hydrolyzed to gluconic acid and H2O2 by glucose oxidase (GOx) (Witt et al., 2000). Most electrochemical glucose sensors rely on H2O2 to monitor glucose; however, we considered that it would be advantageous to use both H2O2 and gluconic acid, as this should theoretically improve sensitivity. Since both molecules are acidic, the enzymatic reaction would be expected to increase the acidity of the solution. In addition, carboxylic acid-doped LCs yield a rapid response to small pH changes (Bi et al., 2009). In general, at a higher pH value, a dark image will be observed, while a lower pH will result in a bright texture. Moreover, when 5CB is exposed to UV light, a photochemical reaction occurs to produce 4-cyano-40 -biphenylcarboxylic acid (CBA). These suggest that UV-treated 5CB should exhibit pH-dependent optical signals. When UV-treated 5CB confined to a gold grid was immersed in a phosphate-buffered saline (PBS) solution (pH 6.0), a bright image with good fan-shape was observed under a cross-polarized microscope (Fig. 1A). We still observed a bright image even after we increased the pH value to 7.4 (Fig. 1B). Interestingly, a slight increase to pH 7.5 immediately resulted in a dark image (Fig. 1C). After increasing the pH to 9.0, the image remained dark (Fig. 1D). In conclusion, a nominal change of pH from 7.4 to 7.5 triggered a bright-to-dark image transition. It is well known that planar orientations of LCs at the aqueous/ LC interface show bright images under a polarized microscope, whereas dark images are caused by homeotropic orientations. Moreover, previous studies have shown that when the adsorption of surfactants at the aqueous/LC interface exceeds a critical value, a planar-to-homeotropic orientational transition occurs (Bi et al., 2009). In our case, the density of CBA and 4-cyano-40 -biphenylcarboxylate (CB) at the aqueous/LC interface correlated with the pH value due to the ionic equilibrium of CBA. Increasing the pH will shift the equilibrium to the right and increase the density of CB but decrease the density of CBA. By contrast, decreasing the pH will shift the equilibrium to the left and decrease the density of CB, while increasing the density of CBA. Due to the amphiphilic nature
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
295
Fig. 1. Polarized microscopy images of UV-treated 5CB confined to gold grids were subjected to different pH PBS solutions: (A) pH ¼ 6.0, (B) pH¼ 7.4, (C) pH¼ 7.5, and (D) pH ¼9.0.
Fig. 2. (a) The schematic setup of the reported glucose sensor; (b) the sensing mechanism. With an unmodified gold grid, the pH of the aqueous solution was maintained at 7.5, sufficient surfactants cause a homeotropic orientation of UV-treated 5CB, resulting in a dark image. With the GOx-modified gold grid, the H þ released from the oxidation of β-D-glucose by GOx decreased the pH of the aqueous solution, a bright image results from carboxylate protonation and a homeotropic-to-planar orientation transition triggered by insufficient surfactants.
of CB and the hydrophobic nature of CBA (40 -Cyano-4biphenylcarboxylic acid | C14H9NO2 | ChemSpider, n.d.), increasing the density of CB (i.e., higher pH) will help introduce a homeotropic orientation and yield a dark image. By contrast, decreasing its density (i.e., lower pH) will facilitate a planar orientation and yield a bright image. As previously alluded to, the critical orientational transition corresponds to a specific pH value between 7.4 and 7.5 (Fig. 2b).
3.2. Monitoring glucose using an enzyme-modified gold grid The GOx-catalyzed oxidation of β-D-glucose is known to release H þ , thereby causing a slight increase in acidity that could be detected by UV-treated 5CB. Thus, we predicted that UV-treated
5CB could be used to monitor the presence of glucose. We constructed the detection system, as shown in the schematic of Fig. 2a, to verify our hypothesis. We first immobilized GOx on a gold grid via NHS/EDC conjugation chemistry (Lahiri et al., 1999). Then, we filled the meshes of the gold grid with UV-treated 5CB. Finally, the whole system was immersed in glucose solution (while carefully adjusting its pH to 7.5 with NaOH and HCl). After the enzymatic reaction was incubated at 40 1C for 2 h, the system was observed with a polarized microscope. The monitoring mechanism is illustrated in Fig. 2b. The H þ released from the oxidation of β-Dglucose by GOx triggered an optical response of the LCs from dark to bright. Fig. 3a-A shows that that the image of the LCs became bright in the presence of 10 pM glucose (pH 7.5). For comparison, the same system was immersed in pure PBS solution (pH 7.5) without glucose. Fig. 3a-B shows that the image remained dark.
296
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
Fig. 3. (a) Polarized light microscopy images of UV-treated 5CB were submerged in (A) glucose solution (pH 7.5) with a GOx-modified gold grid, (B) PBS solution (pH 7.5) with a GOx-modified gold grid, and (C) glucose solution (pH 7.5) with an unmodified gold grid. (b) Images of UV-treated 5CB were submerged in different clinical concentrations of glucose solutions (pH 7.5) after 20 min incubation at 40 1C, (A) 40 mM, (B) 10 mM, (C) 1 mM, and (D) 0.1 mM.
Fig. 4. Polarized light microscopy images of UV-treated 5CB submerged in a series of glucose solutions (pH 7.5) of varying concentrations: (A and G) 50 nM, (B and H) 1 nM, (C and I) 0.1 nM, (D and J) 10 pM, (E and K) 1 pM, and (F and L) 0.1 pM. The gold grids of A–F were modified by GOx, while those of G–L were unmodified.
Furthermore, we used an unmodified gold grid instead of the GOxmodified grid and immersed it in the same glucose solution; Fig. 3a-C shows the dark image obtained. In order to demonstrate the potential of this sensor for fast glucose detection in clinical diagnostics, we prepared a series of glucose solutions (pH 7.5) for which the concentration ranged from 0.1 mM to 40 mM. After 20 min incubation at 40 1C, bright images were obtained under a polarized microscope, as shown in Fig. 3b. These results suggest that this novel system can be used to monitor glucose owing to the H þ released during the enzymatic reaction.
3.3. Detection limit After confirming that glucose could be detected by UV-treated 5CB, we shifted our interest to the detection limit. We prepared a series of glucose solutions (pH 7.5), each with a different concentration. Subsequently, the UV-treated 5CB with a GOx-modified gold grid was immersed in each of these solutions. For comparison, we immersed the UV-treated 5CB with unmodified gold grids into these glucose solutions at the same time (Fig. 4). The images at the top (A–F) of Fig. 4 show the optical responses of UV-treated
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
297
Fig. 5. (a) Histograms of polarized light microscopy images of UV-treated 5CB with GOx-modified gold grids that were in contact with a series of glucose solutions (pH 7.5) of various concentrations. (b) Brightness value as a function of glucose concentration.
298
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
Fig. 6. Polarized microscopy images of UV-treated 5CB confined in the GOx-modified gold grids when placed in contact with different solutions (pH 7.5, 1 mM): (A) sucrose, (B) D-( )-lactic acid, (C) L-(þ )-lactic acid, (D) SDS, (E) L-ascorbic acid, and (F) glucose.
5CB confined in GOx-modified gold grids immersed in glucose solution, and the images at the bottom (G–L) show the responses of unmodified grids immersed in glucose solution. The following concentrations were used: 50 nM (images A and G), 1 nM (B and H), 0.1 nM (C and I), 10 pM (D and J), 1 pM (E and K), and 0.1 pM (F and L). The figure clearly indicates that the concentrations used in A–E produced bright images with good fan shape, while the others produced dark images without fan shape. This result suggests that the detection limit of this system is 1 pM. 3.4. Quantitative measurement of the concentration of glucose solution After identifying the detection limit of our system, we devoted our resources to developing an easier method to measure even lower glucose concentrations. It seems logical that lower glucose concentrations would yield darker images: the less H þ released during the enzymatic reaction means more surfactant (CB) at the aqueous/LC interface, resulting in darker images. Therefore, the analysis of image brightness is a potential method for measuring low glucose concentrations. We plotted a graph of the mean value (in color the term value means brightness) of the images against the logarithm of glucose solution concentration ranged from 1 pM to 50 nM. As shown in Fig. 5b, lower concentrations corresponded to lower brightness values (all the color information was analyzed using GIMP; see Fig. 5a for more details). The brightness of the images could be affected by many factors, including the thickness of the LCs, the depth and pH of the aqueous solution, the intensity of the incident light, and the structure of gold grids. Fortunately, industrial programmed processes can be employed to circumvent (or at least mitigate) these potentially problematic issues. Thus, this method could theoretically be used for measuring low concentrations of glucose with an acceptable degree of accuracy. Herein, we minimized the variation by fixing the illumination light intensity knob with tape, keeping the depth of aqueous solutions at 5 mm while adjusting the pH of all the solutions to 7.5. 3.5. Studying the selectivity of this novel sensor In addition to its extremely high sensitivity, this glucose sensor also displays high selectivity. We were able to exploit the fact that glucose is a type of saccharide, that it exhibits optical activity, that high acid content would affect the blood glucose sensing, and that the surfactant would affect the orientation of the LCs. For example, to demonstrate device selectivity, we examined the optical response of this glucose sensor when immersed in sucrose solution (Fig. 6A, a dark image), D-( )-lactic acid solution (Fig. 6B, a dark image), L-( þ)-lactic acid solution (Fig. 6C, a dark image), sodium dodecyl sulfate (SDS) solution (Fig. 6D, a dark image), and L-ascorbic acid solution (Fig. 6E, a dark image). The concentration of these solutions was 1 mM and the incubation time was 2 h. Fig. 6F shows that a bright image was produced when the sensor was immersed in glucose solution. These results suggest that our
sensor displays high selectivity due to the underlying enzymatic reaction it relies upon.
4. Conclusions In conclusion, we developed a novel glucose sensor based on UV-treated LCs. This sensor can be used not only for clinical glucose detection, but also to monitor low glucose concentrations. It shows outstanding performance with extremely high sensitivity and selectivity. The detection limit of 1 pM is 3 orders of magnitude better than the second most sensitive glucose sensor. These results hold great promise for developing a novel sensor for detecting trace quantities of glucose occurring in biological environments. Furthermore, these UV-treated liquid crystals could be used to monitor other molecules related to pH-changing bioreactions with high sensitivity. The ability to detect trace quantities of many biomolecules will greatly promote the development of several fields of research. Acknowledgment This study was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF2013R1A1A1A05008333) and a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (HI13C0891). References 40 -Cyano-4-biphenylcarboxylic acid | C14H9NO2 | ChemSpider, n.d. [WWW Document]. URL 〈http://www.chemspider.com/Chemical-Structure.2038713.html〉. Aldred, M.P., Contoret, A.E.A., Farrar, S.R., Kelly, S.M., Mathieson, D., O'Neill, M., Tsoi, W.C., Vlachos, P., 2005. Adv. Mater. 17, 1368–1372. Aslan, K., Lakowicz, J.R., Geddes, C.D., 2004. Anal. Biochem. 330, 145–155. Bankar, S.B., Bule, M.V., Singhal, R.S., Ananthanarayan, L., 2009. Biotechnol. Adv. 27, 489–501. Bi, X., Hartono, D., Yang, K.-L., 2009. Adv. Funct. Mater. 19, 3760–3765. Chen, T., Friedman, K.A., Lei, I., Heller, A., 2000. Anal. Chem. 72, 3757–3763. Chen, W., Yao, H., Tzang, C.H., Zhu, J., Yang, M., Lee, S.-T., 2006. Appl. Phys. Lett. 88, 213104. Coles, H., Morris, S., 2010. Nat. Photonics 4, 676–685. He, F., Tang, Y., Yu, M., Wang, S., Li, Y., Zhu, D., 2006. Adv. Funct. Mater. 16, 91–94. Hu, Q.-Z., Jang, C.-H., 2011. Colloids Surf. B Biointerfaces 88, 622–626. Hu, Q.-Z., Jang, C.-H., 2012. J. Biotechnol. 157, 223–227. James, T.D., Harada, T., Shinkai, S., 1993. J. Chem. Soc. Chem. Commun. 10, 857–860. James, T.D., Sandanayake, K.R.A., Shinkai, S., 1996. Angew. Chem. Int. Ed. Engl. 35, 1910–1922. Kinsinger, M.I., Sun, B., Abbott, N.L., Lynn, D.M., 2007. Adv. Mater. 19, 4208–4212. Lahiri, J., Isaacs, L., Tien, J., Whitesides, G.M., 1999. Anal. Chem. 71, 777–790. Lee, G., Carlton, R.J., Araoka, F., Abbott, N.L., Takezoe, H., 2013. Adv. Mater. 25, 245–249. Liao, C., Zhang, M., Niu, L., Zheng, Z., Yan, F., 2013. J. Mater. Chem. B 1, 3820–3829. Lin, I.-H., Miller, D.S., Bertics, P.J., Murphy, C.J., de Pablo, J.J., Abbott, N.L., 2011. Science 332, 1297–1300. Liu, H., Qian, X., Wang, S., Li, Y., Song, Y., Zhu, D., 2009. Nanoscale Res. Lett. 4, 1141–1145. Liu, Y., Cheng, D., Lin, I.-H., Abbott, N.L., Jiang, H., 2012. Lab Chip 12, 3746–3753. Lo, Y.-L., Yu, T.-C., 2006. Opt. Commun. 259, 40–48.
S. Zhong, C.-H. Jang / Biosensors and Bioelectronics 59 (2014) 293–299
Meadows, D.L., Schultz, J.S., 1993. Anal. Chim. Acta 280, 21–30. Park, J.-S., Jang, C.-H., Tingey, M.L., Lowe, A.M., Abbott, N.L., 2006. J. Colloid Interface Sci. 304, 459–473. Rong, L.-Q., Yang, C., Qian, Q.-Y., Xia, X.-H., 2007. Talanta 72, 819–824. Rowinski, P., Rowinska, M., Heller, A., 2008. Anal. Chem. 80, 1746–1755.
299
Sapre, A.G., Bedekar, A., Deshpande, A.V., Lali, A.M., 2000. Biotechnol. Lett. 22, 569–573. Tang, B., Cao, L., Xu, K., Zhuo, L., Ge, J., Li, Q., Yu, L., 2008. Chemistry 14, 3637–3644. Terry, L.A., White, S.F., Tigwell, L.J., 2005. J. Agric. Food Chem. 53, 1309–1316. Witt, S., Wohlfahrt, G., Schomburg, D., Hecht, H.-J., Kalisz, H.M., 2000. Biochem. J. 347, 553–559.
