Discrimination of Umami Tastants Using Floating Electrode-Based

Nov 12, 2015 - ... Hyun Seok Song , Minju Lee , Seunghun Hong , and Tai Hyun Park. ACS Nano 2017 11 (12), 11847-11855. Abstract | Full Text HTML | PDF...
1 downloads 0 Views 3MB Size
Download PDF
ARTICLE

Discrimination of Umami Tastants Using Floating Electrode-Based Bioelectronic Tongue Mimicking Insect Taste Systems Minju Lee,†,^ Je Won Jung,‡,^ Daesan Kim,§ Young-Joon Ahn,‡ Seunghun Hong,*,† and Hyung Wook Kwon*,‡ †

Department of Physics and Astronomy, and Institute of Applied Physics, Seoul National University, Seoul 151-747, Korea, ‡Biomodulation Major, Department of Agricultural Biotechnology and Research Institute of Agriculture and Life Sciences, College of Agriculture & Life Sciences, Seoul National University, Seoul 151-921, Korea, and §Department of Biophysics and Chemical Biology, Seoul National University, Seoul 151-747, Korea. ^These authors contributed equally to this work.

ABSTRACT We report a floating electrode-based bioelectronic

tongue mimicking insect taste systems for the detection and discrimination of umami substances. Here, carbon nanotube fieldeffect transistors with floating electrodes were hybridized with nanovesicles containing honeybee umami taste receptor, gustatory receptor 10 of Apis mellifera (AmGr10). This strategy enables us to discriminate between L-monosodium glutamate (MSG), best-known umami tastant, and non-umami substances with a high sensitivity and selectivity. It could also be utilized for the detection of MSG in liquid food such as chicken stock. Moreover, we demonstrated the synergism between MSG and disodium 50 -inosinate (IMP) for the umami taste using this platform. This floating electrode-based bioelectronic tongue mimicking insect taste systems can be a powerful platform for various applications such as food screening, and it also can provide valuable insights on insect taste systems. KEYWORDS: umami taste . carbon nanotube . nanovesicle . honeybee umami taste receptor . bioelectronic tongue

T

he sense of taste is closely associated with a human life, because it provides important information about the food quality, such as nutritious rich substrates promoting feeding and harmful substrates inhibiting feeding.1,2 In human and some animal species, sweet, sour, salty, and bitter tastes play an important role in discriminating basic taste qualities. Besides basic taste senses, umami taste, which represents the taste sense of amino acids such as glutamate and 50 -ribonucleotides including inosinate and guanylate, was first recognized as a basic taste sense in 1908.3 L-Monosodium glutamate (MSG) and 50 -ribonucleotides are known to elicit the umami taste via umami receptor in human and mammals.4 These taste components are important for food palatability and acceptance in human. These are widely present in meat, fish, and mushrooms, as well as other food products, but the excessive exposure to MSG could LEE ET AL.

cause health problems such as numbness, weakness, and so on.58 Thus, it can be very important to detect excessive MSG in food. A honeybee, which has been a central insect research model for the study of chemosensory perception9 and learning and memory10 can detect a repertoire of taste qualities similar to humans.11 Although the honeybee genome project has revealed 10 gustatory receptor genes,12 the ligands for these gustatory receptors have not yet been identified experimentally. Recently, gustatory receptor 1 of Apis mellifera (AmGr1) was identified as a sugar receptor in the honeybee13 and gustatory receptor 10 of A. mellifera (AmGr10) was specifically tuned to small sets of L-amino acids including umami tastants such as glutamic acid and aspartic acid (unpublished data). Here, we employed a floating electrode-based bioelectronic sensor to utilize the chemosensory function of VOL. 9

’

NO. 12

’

* Address correspondence to [email protected], [email protected]. Received for review May 19, 2015 and accepted November 12, 2015. Published online November 12, 2015 10.1021/acsnano.5b03031 C 2015 American Chemical Society

11728–11736

’

2015

11728 www.acsnano.org

LEE ET AL.

ARTICLE

AmGr10, and investigate if AmGr10 could respond to umami substances selectively. The most noticeable feature of the umami taste is a synergism between MSG and 50 -ribonucleotides such as disodium 50 -inosinate (IMP). 50 -Ribonucleotides are well-known as enhancers of umami taste.14 This indicates that 50 -ribonucleotides decrease the threshold of sensitivity to MSG and thus increase the sensitivity to the umami taste.4,15 Recently, there has been a large amount of research on the development of various MSG sensors based on nanomaterials.6,16 However, these methods intermittently suffered from several limitations. For instance, previous research using carbon nanotubes functionalized with glutamate oxidase can only be employed to detect glutamatebased umami tastants such as MSG.16 Taken together, these conventional methods have limitations to characterize the synergism that is the hallmark of the umami taste. Technically, the performance of sensor transducers is a critical issue in bioelectronic sensors, because it significantly affects device sensitivity and selectivity. Previously, it has been reported that the combination of a floating electrode structure and a carbon nanotube field-effect transistor (CNT-FET) exhibited an enhanced performance as a sensor transducer compared with conventional CNT-FETs.17 However, floating electrode structures have not been employed for bioelectronic sensor applications, and previous work in this field has relied on simple FET structures.18,19 Herein, we report a floating electrode-based bioelectronic tongue mimicking insect taste systems for the discrimination of umami taste substances. In this work, nanovesicles containing honeybee umami taste receptor, AmGr10, were immobilized on a floating electrode-based carbon nanotube sensor transducer to serve as a bioelectronic tongue device for the detection of specific umami tastants. This sensor could discriminate MSG, well-known umami tastant, from non-umami substances with a high sensitivity and selectivity. Also, we could detect umami tastants directly in liquid food such as chicken stock. Significantly, we could quantitatively evaluate the synergistic effects of IMP to enhance the detection of MSG using a floating electrode-based carbon nanotube sensor incorporated with nanovesicles containing AmGr10. Since this method relies on receptor molecules instead of enzyme, it can be utilized to detect a broad range of umami tastants without being limited to glutamatebased ones. Furthermore, the floating electrodebased devices allow one to improve the sensitivity of previous CNT-based sensors. Our method can be an effective strategy as an artificial taste sensor, and thus provides the broad opportunities for basic research on taste sensory systems of animals toward various practical applications in food and other related industries.

Figure 1. Schematic diagram depicting the fabrication process of a floating electrode-based bioelectronic tongue mimicking insect taste systems. A CNT-FET with floating electrodes was hybridized with nanovesicles containing honeybee umami taste receptor, AmGr10. The CNT-FET with floating electrodes was fabricated by photolithography processes. The nanovesicles containing AmGr10 functionalized with thiol groups were immobilized on the gold floating electrodes. AmGr10 can respond to specific umami tastants with a high selectivity.

RESULTS AND DISCUSSION Figure 1 shows a schematic diagram describing a floating electrode-based carbon nanotube sensor transducer hybridized with nanovesicles containing an umami taste receptor of western honeybees, AmGr10. First, we fabricated a CNT-FET with floating electrodes as described previously.2023 Briefly, CNTs were selectively adsorbed on SiO2 substrate, followed by fabrication of the metal electrodes by thermal deposition. Finally, the source and drain electrodes were covered by a passivation layer. It has been reported that a biosensor based on a CNT-FET with floating electrodes exhibited higher sensor signals than a conventional biosensor based on a simple CNT-FET.23 Following the fabrication of the CNT-FET with floating electrodes, we immobilized the nanovesicles containing AmGr10 on the floating electrodes of the device. The AmGr10, umami taste receptor, could recognize specific umami tastants. Our fabrication method allows us to mass-produce devices in a wafer scale. In this work, we could fabricate 24 devices on a single wafer and performed the sensing experiments using four or more devices for each data point to confirm the reliability of our method. Figure 2a is the field emission scanning electron microscopy (FE-SEM) image of nanovesicles immobilized on a gold surface. Before the SEM imaging, the nanovesicles were lyophilized to maintain their structures and then coated with 10 nm thick platinum using a sputter coater. This image shows that the nanovesicles could be immobilized uniformly on solid surfaces. Figure 2b shows Western blot analysis for the confirmation of AmGr10 expression. In brief, the expressions of AmGr10 protein in HEK-293T cells and cellderived nanovesicles were confirmed from the lysates VOL. 9

’

NO. 12

’

11728–11736

’

11729

2015 www.acsnano.org

ARTICLE Figure 2. Expression of AmGr10 in HEK-293T cells and nanovesicles. (a) FE-SEM image of nanovesicles containing AmGr10 on the gold substrate. (b) Western blot analysis of AmGr10 protein expression in HEK-293T cells and cell-derived nanovesicles. We have obtained rat antiserum against a synthetic peptide from the sequence of AmGr10. Transfected cells and nanovesicles with AmGr10 exhibited the specific band corresponding to the molecular weight of AmGr10, while the control cells and nanovesicles (without AmGr10) did not show the band. (c) Dose-dependent intracellular Ca2þ assay in HEK-293T cells expressing AmGr10 upon the addition of MSG. Only the addition of MSG caused Ca2þ influx in the HEK-293T cells. Each point represents the mean ( SEM of at least 10 assays. One-way ANOVA test followed by Bonferroni correction for multiple comparison was employed to test the difference in dose-dependent responses of MSG (p < 0.001), while Student's t-test was used to examine the different responses between MSG and other substances (p < 0.001). (d) Real-time measurement of Ca2þ assay in the nanovesicles containing AmGr10. The addition of 10 mM of MSG resulted in the increase of fluorescence intensity compared with control nanovesicles without receptors. Each point represents the mean ( SEM (Student's t-test: p < 0.05, N = 10).

of transfected with HEK-293T cells and nanovesicles by Western blot analysis. The AmGr10 antibody is polyclonal and was obtained from rats exposed to a synthetic peptide from the AmGr10 sequence (see Materials and Methods). Lane 1 represents the data from AmGr10-expressing cells and nanovesicles derived from them, while lane 2 represents the data from control cells and nanovesicles. The band of 42 kDa which shows the molecular weight of AmGr10 was observed from the AmGr10-expressing cells and cellderived nanovesicles, while the band was not observed from control cells and nanovesicles. These data indicate that AmGr10, the umami taste receptor of the honeybees, was expressed in HEK-293T cells and nanovesicles. In this manner, we determined that the cell derived-nanovesicles contained a sufficient amount of the umami taste receptor of the honeybees. To investigate the functional activity of AmGr10, we carried out the measurement of intracellular calcium concentration changes in AmGr10-expressing HEK-293T cells upon the stimulation of the umami tastant, MSG. Figure 2c shows the dose-dependent intracellular Ca2þ assay using Fluo-4 in HEK-293T cells expressing AmGr10. The Ca2þ responses were measured by calcium indicator Fluo-4 using a spectrofluorophotometer (see Materials and Methods). The AmGr10-expressing HEK-293T cells exhibited Ca2þ responses to MSG above a threshold concentration of 30 mM in a dosedependent manner. However, intracellular signal transduction did not occur by sweeteners in transfected cells LEE ET AL.

with AmGr10. These results clearly show that the AmGr10 was well expressed in the cells, and it selectively responded to umami tastants such as MSG. Also, we performed a calcium image analysis to identify whether AmGr10-mediated Ca2þ influx could activate calcium signals in nanovesicles. Figure 2d shows the real-time measurement of Ca2þ assay in the nanovesicles containing AmGr10. Note that treatment of nanovesicles containing AmGr10 with MSG (10 mM) resulted in the immediate increase of fluorescence ratio compared to the control nanovesicles without AmGr10. It indicates that the binding of MSG onto the AmGr10 induced a Ca2þ influx into the nanovesicles. These data suggested that functional AmGr10 was expressed and incorporated into the cell's plasma membrane, and retained its functional response to agonist when isolated in nanovesicles. In this case, the recovery of calcium signaling to the baseline in nanovesicles was not observed, probably because of the lack of ion pumps and calmodulin which are necessary to restore the Ca2þ concentration.19,24 Figure 3a shows the real-time response to various concentrations of MSG obtained by a floating electrode-based bioelectronic tongue. A bias voltage of 0.1 V was applied and maintained during electrical measurements. Here, source-drain currents were monitored after the introduction of MSG solutions to the device. As shown in Figure 3a, the introduction of MSG solutions resulted in the increase of FET channel conductance with a dose-dependent manner. Here, VOL. 9

’

NO. 12

’

11728–11736

’

11730

2015 www.acsnano.org

ARTICLE Figure 3. Detection of MSG with floating electrode-based bioelectronic tongues. (a) Real-time electrical measurement of MSG. The introduction of MSG caused an increase in the FET channel conductance. The FET channel conductance began to increase after the addition of MSG with 100 pM. (b) Dose-dependent responses of floating electrode-based bioelectronic tongues to MSG. The normalized signal increased as the concentration of MSG increased, and it was saturated at around 10 μM of MSG. We repeated sensing measurements for four or more bioelectronic tongue devices to confirm the reliability. The error bars represent the standard deviations of the normalized sensor signals. (c) Real-time response of a floating electrode-based bioelectronic tongue to various tastants. The non-umami tastants (sucrose and PTC) caused negligible FET channel conductance changes, while the introduction of MSG resulted in large increase in FET channel conductance. (d) Realtime response of a floating electrode-based bioelectronic tongue to the commercial chicken stock solutions diluted with DPBS solutions by 1/106 and 1/105 times from as-purchased commercial chicken stock.

the sensor signal (|ΔG/G0|) represents relative FET channel conductance change at certain concentration. Our sensor began to respond to MSG at a concentration of 100 pM, indicating that the floating electrode-based bioelectronic tongue could respond to MSG in real time with a high sensitivity. A plausible explanation is that the binding of MSG onto the AmGr10 induced a Ca2þ influx into the nanovesicles. Subsequently, the increased concentration of Ca2þ in nanovesicles resulted in the increase of the FET channel conductance via the modulation of the Schottky barrier between the CNT networks and the floating electrodes.23 Figure 3b shows the normalized signals of floating electrode-based bioelectronic tongues at various concentrations of MSG. The normalized signals of floating electrode-based bioelectronic tongues to each tastant were obtained by the normalization of the sensor signals with respect to their maximum signal values at high concentrations.18 We repeated the sensing measurements for four or more bioelectronic tongue devices to calculate the averaged values and standard deviations. Even though we used devices fabricated under same conditions, there is a little bit of a difference in device characteristics, which led to the variation in responses. Rather narrow error bars indicate that we could obtain reproducible and reliable responses from our floating electrode-based bioelectronic tongues. LEE ET AL.

At a 100 pM or higher concentrations, the sensors exhibited normalized signals larger than error bars, indicating that our sensor can detect MSG down to 100 pM. The normalized signal increased as the concentration of MSG increased, and it was saturated at a high concentration of 10 μM. Here, the dose-dependent responses of the floating electrode-based bioelectronic tongues can be analyzed by the model based on the Hill equation as reported previously.18,19,25,26 First, we assume that the binding events between receptors (AmGr10) and MSG follow the Hill equation. Then, the density Cs of MSG bound to the receptors can be written as Cs ¼

Cs, max 3 C n (1=K)n þ C n

(1)

Here, C and K are the concentration of MSG in a solution and the equilibrium constant between the AmGr10 and MSG, respectively. Cs,max is the density of AmGr10 on the floating electrodes, and n is the value of the Hill coefficient. If we assume that a conductance change ΔG is linearly proportional to the number of bound MSG, the sensor signal |ΔG/G0| can be approximated as |ΔG/G0| ∼ kCs, where k is a constant representing the response characteristics of the floating electrode-based bioelectronic tongue. When C becomes VOL. 9

’

NO. 12

’

11728–11736

’

11731

2015 www.acsnano.org

N ¼

Cn (1=K)n þ C n

(2)

By fitting the experimental data using eq 2, we can estimate the equilibrium constant K between AmGr10 and MSG as 1.77  108 M1. This quantitative analysis also helps us to predict the responses of our floating electrode-based bioelectronic tongues with the umami taste receptor of the honeybees to its ligand. Figure 3c shows the real-time response of the floating electrode-based bioelectronic tongue to various tastants. Sucrose and phenylthiocarbamide (PTC) are sweet and bitter taste compounds, respectively. The injection of the sucrose and PTC with a high concentration of 100 nM caused negligible conductance changes, while that of low concentration MSG caused a sharp increase in the FET channel conductance. Also, the addition of another sweetener, glucose (100 nM), did not result in the conductance change (Figure S1 in Supporting Information). This result implies that the floating electrode-based bioelectronic tongue discriminates umami tastants from non-umami tastants with a high selectivity. To demonstrate the capability of the floating electrode-based bioelectronic tongue for practical applications, we also performed experiments to detect MSG in a food sample, a commercial chicken stock (Figure 3d). The detailed experimental procedure is described in the Materials and Methods section. In brief, the concentration of glutamic acid in the commercial chicken stock was first measured via the high performance liquid chromatography (HPLC) method (Figure S2 in Supporting Information),27 and the chicken stock was mixed with Dulbecco's Phosphate Buffered Saline (DPBS) solutions with different ratios to prepare watery chicken stock solutions. On the basis of the HPLC measurement results, the concentrations of glutamic acids in chicken stock solutions diluted by 1/106 and 1/105 times could be estimated as 7.1  1010 and 7.1  109 M, respectively. Then, the diluted chicken stock solutions were applied to a bioelectronic tongue while monitoring its responses (Figure 3d). The addition of the chicken stock solution caused a significant increase in the FET channel conductance, which shows that our floating electrode-based bioelectronic tongues can detect umami tastants in liquid food. By fitting the sensor responses via the Hill equation obtained from the data in clean DPBS solution environments (Figure 3b), we could estimate the glutamic acid concentrations of 6.2  1010 and 4.9  109 M in the chicken stock after the addition of 1/106 and 1/105 times diluted chicken stock solutions, respectively. These values are close to the glutamic acid concentrations estimated by the HPLC method within the error LEE ET AL.

bars in Figure 3b. These results clearly show that the floating electrode-based bioelectronic tongue can detect MSG in complicated environments such as chicken stock as well as DPBS solution and it can be utilized as a sensor platform for practical applications. Furthermore, we investigated the synergistic effects of disodium 50 -inosinate (IMP), which is known as an enhancer of umami taste, on detecting umami tastant (MSG). This characteristic is a distinctive feature of umami taste sensory systems.1,28,29 Figure 4a shows the Ca2þ fluorescence assay images showing the dosedependent responses of AmGr10 in HEK-293T cells to MSG alone and the mixture of MSG and IMP. To examine the effect of IMP, HEK-293T cells expressing AmGr10 were stimulated with various concentrations of MSG in the presence and absence of 2.5 mM (98 mg/100 g) IMP. Previous reports show that IMP can exist in foods with its concentration ranging from 0 to 474 mg/100 g and can give a synergistic effect.3033 Our results show that HEK-293T cells expressing AmGr10 responded to MSG at a concentration of 100 mM in the absence of IMP (Figure 4a). In contrast, in the presence of 2.5 mM IMP, the responses of HEK293T cells expressing AmGr10 to MSG presented much higher fluorescent intensity than the responses in the absence of IMP. Additionally, IMP alone did not activate AmGr10. This result indicates that the response to umami tastant can be strongly enhanced by IMP. This result allows us to verify the synergism, the hallmark of umami taste systems. Figure 4b shows doseresponse curves of AmGr10 to MSG in the presence and absence of 2.5 mM IMP based on quantitative analysis of Ca2þ imaging in Figure 4a. We repeated the experiments with nine or more cell samples for quantitative analysis. The experimental data obtained from Figure 4a were fitted by the Hill equation to evaluate equilibrium constants. Here, we could estimate K, which is the equilibrium constant between AmGr10 and MSG, as 3.45  10 M1. Likewise, the equilibrium constant Ks by the synergism between MSG and IMP can be estimated as 7.14  102 M1. These equilibrium constants can be converted to EC50 values. The evaluated EC50 values are 78.76 and 1.73 mM, respectively. This result shows that the EC50 value of AmGr10 to MSG was shifted 45-fold in the presence 2.5 mM IMP. This result implies that the responses of AmGr10 to MSG were strongly potentiated by concentrations of IMP comparable to what is found in some foods. Previous studies have shown the molecular mechanism of synergistic effect of IMP. Human umami taste receptors, T1R1 and T1R3, possess extracellular Venus flytrap domain (VFTD) that consists of two lobes.14 L-Glutamate binding site of T1R1/T1R3 lies a hinge region of VFTD of T1R1 and IMP binds to an adjacent site close to the opening region of the VFTD of T1R1 using site-directed mutagenesis and molecular modeling.14,34 Therefore, IMP may act in the extracellular domain of AmGr10 in a similar manner VOL. 9

’

NO. 12

’

11728–11736

’

ARTICLE

very large, the sensor signal |ΔG/G0| converges to the value of kCs,max. Then, we could write the normalized signal N as follows:

11732

2015 www.acsnano.org

ARTICLE Figure 4. Synergism between MSG and IMP in HEK-293T cells and cell-derived nanovesicles. (a) Ca2þ fluorescence assay showing doseresponses of AmGr10 to MSG and IMP in HEK-293T cells. The HEK-293T cells expressing AmGr10 were stimulated by MSG with and without IMP (2.5 mM). The results indicated that AmGr10 was activated by MSG and the responses were enhanced by IMP. (b) Doseresponse curves of AmGr10 in HEK-293T cells to MSG and MSG with IMP based on quantitative analysis of Ca2þ imaging. We repeated the experiments with nine or more cell samples for quantitative analysis. The error bars represent the standard deviations of the normalized responses. The curves were fitted to the Hill equation. The EC50 value of AmGr10 to MSG was shifted 45-fold in the presence 2.5 mM IMP. Values depict the mean ( SEM. (c) Normalized signals of floating electrode-based bioelectronic tongues with nanovesicles in the presence and absence of 2.5 nM IMP. We repeated sensing measurements for four or more bioelectronic tongue devices to confirm the reliability. The error bars represent the standard deviations of the normalized sensor signals. By fitting the curves using the Hill equation, we estimated equilibrium constant K between AmGr10 and MSG as 1.77  108 M1 and the equilibrium constant Ks with synergism between MSG and IMP as 2.30  109 M1. The result showed that the responses to MSG could be enhanced by the synergism between MSG and IMP.

and can strongly potentiate the umami taste intensity as an enhancer in insect taste systems. We also examined the synergism by utilizing floating electrode-based bioelectronic tongues. Figure 4c shows the normalized signals of our platform in the presence and absence of 2.5 nM IMP. Here, the sensing measurements were repeated using four or more bioelectronic tongue devices to confirm the reliability. Overall, the bioelectronic tongues responded to the MSG at much lower concentrations than in the case of cell-based assay in Figure 4b. Previous works show that nanovesicles responded to analyte solutions with much lower concentrations than cells.19,35,36 Such results were attributed to the small volume of nanovesicles compared with cells. For example, as the radius R of a vesicle decreases, the volume (∼R3) decreases much faster than the surface area (∼R2). Note that the number of ion channels on the vesicle surface should depend on the surface area of the vesicle. In this case, as the vesicle size decreases, its volume should decrease much faster than the number of ion channels. Thus, relatively small nanovesicles can be filled up more easily by calcium ions, and they respond to lower concentration analytes than cells with a rather large volume.36 To verify the synergism, we prepared the LEE ET AL.

mixture of MSG and IMP according to the procedure described previously.33 In the mixture, the concentration of IMP was fixed at 2.5 nM, and that of MSG varied from 1011 to 5  108 M. As shown in Figure 4c, the response curve to MSG was shifted toward lower concentrations. We could also analyze the experimental data by using the model based on the Hill equation. The equilibrium constant K between AmGr10 and MSG can be estimated as 1.77  108 M1 in Figure 3b. We could estimate Ks, which is the equilibrium constant by the synergism between MSG and IMP, as 2.30  109 M1 in the same way. This result shows that the response of AmGr10 to MSG was significantly enhanced by the synergism between MSG and IMP. That is, the synergistic effects of IMP decrease the threshold of AmGr10 response to MSG. These results show that we could quantitatively evaluate the synergism which is the hallmark of umami taste using the floating electrode-based bioelectronic tongue. CONCLUSIONS We have successfully developed a floating electrodebased bioelectronic tongue mimicking insect taste systems for the discrimination of umami taste substances using the hybridization of floating electrode-based VOL. 9

’

NO. 12

’

11728–11736

’

11733

2015 www.acsnano.org

methods in terms of the sensitivity and less restrictive experiment conditions. In light of these, the floating electrode-based bioelectronic tongue mimicking insect taste systems can be a simple, but highly effective strategy in many different basic research areas about sensory systems. Moreover, our research provides opportunities to develop various applications in food industry and to research the insect taste systems.

MATERIALS AND METHODS

immediately or could be stored at 80 C for a few weeks for the subsequent experiments. Western Blot Analysis. Rat polyclonal antibodies against AmGr10 were generated by the Abclone company (Seoul, Korea). On the basis of peptide information on AmGr10 (GenBank: NP_001229923.1) and the sequence alignment of peptide, a target epitope region was chosen as NH2-SMNTQILIFVCILFLIEC to produce polyclonal anti-AmGr10 antibody. Rats were immunized three times with 0.5 mg of the synthesized AmGr10 peptide. Serum-specific antibody was purified based on the affinity to immobilized antigen peptides. For Western blotting, transfected cells containing AmGr10 were homogenized in 1 mL of ice-cold 0.1 M Tris-HCl buffer (pH 7.8 with 0.5% Triton X-100 and 1 μL protein inhibitor cocktail). Electrophoresis was carried out with a vertical electrophoresis unit (Novexmini cell, Invitrogen, Carlsbad, CA). Protein preparation from the cell was separated by SDS-PAGE (10%) in triplicate at 130 V for 100 min in a chamber. Proteins separated on the gel were transferred onto a nitrocellulose membrane (Whatman, Pittsburgh, PA) by electroblotting. The membrane was incubated in PBS buffer containing 0.1% Tween-20 and 5% fat-free dry milk for 1 h at room temperature. The primary AmGr10 antibody solution was applied to the blocked membrane sheet, and it was incubated at 4 C overnight. The membrane was then incubated with the horseradish peroxidase conjugated with anti-rabbit IgG secondary antibody for 3 h. The antigen antibody complex was visualized using a chemiluminescence kit according to the manufacturer's instruction (Young In Frontier Co., Seoul, Korea). Intracellular Calcium Assay. For a calcium signaling assay, HEK293T cells expressing AmGr10 were cultured for more than 3 days. Fluo-4 NW dye mix solution (Molecular Probes) was loaded into the cell in 96 well plates. Then, the cells were incubated at 37 C for 1 h. The fluorescence signal upon the addition of MSG (30, 50, 100, 300, and 500 mM) was measured at 516 nm by the excitation at 494 nm using a spectrofluorophotometer (PerkinElmer, Waltham, MA). To normalize the response, the changes of fluorescence ratio were divided by maximal fluorescence changes induced by each compounds. To perform calcium assay of nanovesicles, nanovesicles containing AmGr10 were immobilized on poly-D-lysine-treated 96 well plates by incubation at 37 C for 2 h. The procedure of the calcium assay upon the addition of 10 mM MSG was same for the cells expressing AmGr10 described above. Calcium Imaging. Transfected cell lines were subsequently cultured on confocal dishes (SPL, Pocheon, Korea). Before experiments, the culture media was removed. Then, 100 μL of 2 μM Fluo-4 AM (Invitrogen, Grand Island, NY) was loaded into the cells. The cells were incubated in the dark for 90 min at room temperature. After incubation, 1 HBSS buffer was added onto the confocal dish. The dish was then directly placed in the LSM700 inverted confocal microscope for observation (Zeiss, Oberkochen, Germany). Images were captured with the interval of maximum 100 frames per every 2 s. Test chemicals were dissolved in HBSS at various concentrations. Ca2þ influx into the transfected cells upon ligand binding was monitored and analyzed with ZEN software (Zeiss, Oberkochen, Germany). Fabrication of a CNT-FET with Floating Electrodes. ssCNTs were dispersed in 1,2-dichlorobenzene using ultrasonic cleaner for 5 h. The concentration of ssCNT solution was 0.05 mg/mL.

Materials. Single-walled semiconducting 99% carbon nanotubes (ssCNTs) were purchased from NanoIntegris, Inc. and used as received in our experiments. MSG, IMP, sucrose, glucose, PTC, and other chemical reagents were purchased from SigmaAldrich (St. Louis, MO) and used as received. Honeybee Preparations. Honeybee species, A. mellifera, were maintained in the apiary at Seoul National University campus surrounded by the Gwan-ak mountain range in Seoul, Korea. For cDNA synthesis, worker honeybees were captured directly from three hives regardless of age. They were placed in glass vials and cooled on ice until they stopped moving. HEK-293T Cell Culture. Human embryonic kidney-293T (HEK293T) cells were cultured at 37 C under 5% CO2. Dulbecco's modified Eagle medium (DMEM, Invitrogen, Carlsbad, CA) containing 10% fatal bovine serum and 0.5% penicillinstreptomycin (Invitrogen, Carlsbad, CA) was used to culture HEK-293T cells. RNA Extraction, cDNA Synthesis, and RT-PCR. Total RNAs were extracted from the antennae of honeybee using a Qiagen RNeasy kit according to the manufacturer's instructions (Qiagen, Valencia, CA). Reverse transcription procedures were carried out as described previously.37 Heterologous Expression of AmGr10 into HEK-293T Cells. The expression vector was synthesized by inserting the cDNA of AmGr10 (A. mellifera gustatory receptor 10) into the multiple cloning site of the pcDNA3.1 vector using the restriction enzymes EcoRI and NotI (Koscamco, Anyang, Korea). Template pDNA for pcDNA3.1-AmGr10 and primers were then mixed with kit solutions. The total PCR reaction volume and conditions were followed as described previously.13 One microgram of the vector as a combination AmGr10 was transfected to the HEK-293T cells by using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). The transfection method was according to the manual. In brief, the Opti-MEM (Invitrogen, Carlsbad, CA) was mixed with Lipofectamine at the rate of 2 μL in 50 μL Opti-MEM. It was incubated for 5 min at room temperature. The plasmid DNA (1 μg) was also blended with 50 μL of Opti-MEM media. Two mixed media were combined and incubated for 20 min at room temperature. The solution was applied on the HEK-293T cells. After 24 h, the cell was selected using the concentration of 200 μg/mL of zeocin antibiotics (Invitrogen, Carlsbad, CA). Selected cells were cultured on blended media with 20 μg/mL zeocin. Construction of Nanovesicles from HEK-293T Cell Expressing Honeybee Umami Taste Receptor. Suspended HEK-293T cells expressing honeybee umami taste receptors in DMEM containing cytochalasin B (20 μg/mL, Sigma-Aldrich, St. Louis, MO), which is known as an active reagent for destabilizing cytoskeleton, were incubated at 37 C with 300 rpm agitation. Cells and cell debris were separated using centrifugation at 2000g for 20 min. Nanovesicles were collected by centrifugation at 12000g for 30 min. The nanovesicles were suspended in Phosphate Buffered Saline (PBS) containing a protease inhibitor cocktail (Sigma-Aldrich, St. Louis MO). The production of nanovesicles induced by cytochalasin B did not involve any cell homogenizing steps which may release cytoplasmic proteins or change membrane protein orientation. Therefore, the orientation and functional activity of cell surface receptors were likely to be preserved in vesicles.35 Produced nanovesicles were used

LEE ET AL.

VOL. 9

’

NO. 12

’

11728–11736

’

ARTICLE

CNT-FETs and nanovesicles containing honeybee umami taste receptor, AmGr10. This bioelectronic tongue recognizes MSG down to 100 pM and discriminates between umami and non-umami substances with a high sensitivity and selectivity. Also, the bioelectronic tongue perceives the presence of MSG in liquid food such as chicken stock. Importantly, we have demonstrated the synergism between MSG and IMP. Our strategy overcomes the limitations of previous

11734

2015 www.acsnano.org

Acknowledgment. This work was supported by the NRF Grant (No. 2013M3A6B2078961), World Class University (WCU) program (R31-10056) through the Korea Science and Engineering Foundation funded by the National Research Foundation of Korea, young Investigator grant (112026-1) from Ministry of Agriculture, and Food and Rural Affairs as well as research grant (Project No. PJ010487) from Cooperative Research Program for Agriculture Science & Technology Development from the Rural Development Administration (RDA) of the Republic of Korea. In addition, this work was supported by the BioGreen 21 program of Rural Development Administration (PJ009031, PJ011170) to H.W.K. S.H. also acknowledges the support from the MSIP (Nos. 2013M3C8A3078813, 2014M3A7B4051591). M.L. and D.K. fabricated bioelectronic tongue devices and performed sensing measurements. J.W.J. prepared nanovesicles with honeybee umami taste receptors. H.W.K. and S.H. are responsible for the project and contributed to data analyses. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

LEE ET AL.

Supporting Information Available: The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.5b03031. Figures S1 and S2 (PDF)

REFERENCES AND NOTES 1. Nelson, G.; Chandrshekar, J.; Hoon, M. A.; Feng, L. X.; Zhao, G.; Ryba, N. J. P.; Zuker, C. S. An Amino-Acid Taste Receptor. Nature 2002, 416, 199–202. 2. Yarmolinsky, D. A.; Zuker, C. S.; Ryba, N. J. P. Common Sense about Taste: From Mammals to Insects. Cell 2009, 139, 234–244. 3. Nakamura, E. One Hundred Years since the Discovery of the 00 Umami00 Taste from Seaweed Broth by Kikunae Ikeda, Who Transcended His Time. Chem. - Asian J. 2011, 6, 1659– 1663. 4. Nakashima, K.; Katsukawa, H.; Sasamoto, K.; Ninomiya, Y. Behavioral Taste Similarities and Differences among Monosodium L-Glutamate and Glutamate Receptor Agonists in C57BL Mice. J. Nutr. Sci. Vitaminol. 2001, 47, 161–166. 5. Williams, A. T. R.; Winfield, S. A. Determination of Monosodium Glutamate in Food Using High-Performance Liquid-Chromatography and Fluorescence Detection. Analyst 1982, 107, 1092–1094. 6. Khan, R.; Gorski, W.; Garcia, C. D. Nanomolar Detection of Glutamate at a Biosensor Based on Screen-Printed Electrodes Modified with Carbon Nanotubes. Electroanalysis 2011, 23, 2357–2363. 7. Geha, R. S.; Beiser, A.; Ren, C.; Patterson, R.; Greenberger, P. A.; Grammer, L. C.; Ditto, A. M.; Harris, K. E.; Shaughnessy, M. A.; Yarnold, P. R.; et al. Review of Alleged Reaction to Monosodium Glutamate and Outcome of a Multicenter Double-Blind Placebo-Controlled Study. J. Nutr. 2000, 130, 1058s–1062s. 8. Krishna Veni, N.; Karthika, D.; Surya, D. M.; Rubini, M.; Vishalini, M.; Pradeepa, Y. Analysis of Monosodium L-Glutamate in Food Products by High-Performance Thin Layer Chromatography. J. Young Pharm. 2010, 2, 297–300. 9. Sandoz, J. C. Behavioral and Neurophysiological Study of Olfactory Perception and Learning in Honeybees. Front. Syst. Neurosci. 2011, 5, 98. 10. Giurfa, M.; Sandoz, J. C. Invertebrate Learning and Memory: Fifty Years of Olfactory Conditioning of the Proboscis Extension Response in Honeybees. Learn. Mem. 2012, 19, 54–66. 11. de Brito Sanchez, M. G. Taste Perception in Honey Bees. Chem. Senses 2011, 36, 675–692. 12. Weinstock, G. M.; Robinson, G. E.; Gibbs, R. A.; Worley, K. C.; Evans, J. D.; Maleszka, R.; Robertson, H. M.; Weaver, D. B.; Beye, M.; Bork, P.; et al. Insights into Social Insects from the Genome of the Honeybee Apis Mellifera. Nature 2006, 443, 931–949. 13. Jung, J. W.; Park, K. W.; Ahn, Y. J.; Kwon, H. W. Functional Characterization of Sugar Receptors in the Western Honeybee, Apis Mellifera. J. Asia-Pac. Entomol. 2015, 18, 19–26. 14. Zhang, F.; Klebansky, B.; Fine, R. M.; Xu, H.; Pronin, A.; Liu, H. T.; Tachdjian, C.; Li, X. D. Molecular Mechanism for the Umami Taste Synergism. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 20930–20934. 15. Delay, E. R.; Beaver, A. J.; Wagner, K. A.; Stapleton, J. R.; Harbaugh, J. O.; Catron, K. D.; Roper, S. D. Taste Preference Synergy between Glutamate Receptor Agonists and Inosine Monophosphate in Rats. Chem. Senses 2000, 25, 507–515. 16. Koh, J.; Yi, M.; Yang Lee, B.; Kim, T. H.; Lee, J.; Jhon, Y. M.; Hong, S. Directed Assembly of Carbon Nanotubes on Soft Substrates for Use as a Flexible Biosensor Array. Nanotechnology 2008, 19, 505502. 17. Lee, J.; Lee, H.; Kim, T.; Jin, H. J.; Shin, J.; Shin, Y.; Park, S.; Khang, Y.; Hong, S. Floating Electrode Transistor Based on Purified Semiconducting Carbon Nanotubes for High Source-Drain Voltage Operation. Nanotechnology 2012, 23, 085204.

VOL. 9

’

NO. 12

’

11728–11736

’

ARTICLE

First, an octadecyltrichlorosilane self-assembled monolayer with nonpolar terminal groups was patterned on SiO2 substrate (3000 Å) as reported previously.20,21 For the selective adsorption of CNTs, the substrate was placed in the solution of ssCNTs for 10 s and rinsed with 1,2-dichlorobenzene. Afterward, metal electrodes were fabricated using thermal evaporation (Pd/Au 10 nm/15 nm). The CNT network channel has dimensions of 3 μm width and 170 μm length. Our CNT-FETs have five floating electrodes, and each floating electrode has dimensions of 200 μm width and 10 μm length. Lastly, the source and drain electrodes were covered by a passivation layer (Photoresist, DNR) to eliminate leakage current during the electrical measurements in aqueous environments. Immobilization of Nanovesicles on the Floating Electrodes of the CNT-FET. Nanovesicles containing AmGr10 were dissolved in 2-iminothiolane 3 HCl solution (100 mM in DPBS) for the formation of thiol groups on the surface of nanovesicles, and the 2-iminothiolane 3 HCl solution including the nanovesicles was incubated for 1.5 h at room temperature.3840 Then, the CNTFETs with floating electrodes were incubated in the solution containing nanovesicles for 3.5 h at 4 C in order for the nanovesicles to be selectively adsorbed onto the gold floating electrodes of the CNT-FETs. As a result, nanovesicles with thiol groups were successfully immobilized on the floating electrodes. Preparation of Tastants and Chicken Stock Sample. MSG, IMP, sucrose, glucose, and PTC used in our experiments were dissolved in DPBS solution. A volume of 50 μL of chicken stock solution (Campbell's Soup Company, Camden, NJ), which is available commercially, was mixed with 5 mL of DPBS solution (1% v/v). The mixed chicken stock solution was then filtered through a 0.45-μm-pore-size syringe filter (Advantec, Dublin, CA). Then, the chicken stock solution was mixed with additional DPBS solution to prepare the chicken stock solution diluted by 1/106 and 1/105 times from the as-purchased chicken stock solution. Also, the mixtures of MSG and IMP were prepared to investigate the synergism following the previously reported method.33 Electrical Measurements. For the detection of signals, a CNTFET with floating electrodes was connected to a Keithley 4200 semiconductor analyzer, and a source-drain bias voltage of 0.1 V was maintained during electrical measurements. A 9 μL droplet of DPBS buffer was placed on the channel region of the device, and source-drain currents were monitored in response to the addition of different target materials. In our experiments, ΔG/G0 was defined as the conductance change over the original conductance of the device. HPLC Analysis of Amino Acids. Analysis of amino acids was performed based on the method of previous studies.27 Preparative HPLC (Ultimate3000, Thermo Dionex, Sunnyvale, CA) was used for separation of the constituents from the chicken stock. The column was 4.6  150 mm2 (C18, Waters, VDS optilab, Germany) using acetonitrile/methanol/water (1:4.5:4.5, v/v) at a flow rate of 1.5 mL/min and detected at 338 nm. Conflict of Interest: The authors declare no competing financial interest.

11735

2015 www.acsnano.org

LEE ET AL.

39. Jue, R.; Lambert, J. M.; Pierce, L. R.; Traut, R. R. Addition of Sulfhydryl-Groups to Escherichia-Coli Ribosomes by Protein Modification with 2-Iminothiolane (Methyl 4-Mercaptobutyrimidate). Biochemistry 1978, 17, 5399– 5406. 40. Tarentino, A. L.; Phelan, A. W.; Plummer, T. H. 2-Iminothiolane A Reagent for the Introduction of Sulfhydryl-Groups into Oligosaccharides Derived from Asparagine-Linked Glycans. Glycobiology 1993, 3, 279–285.

VOL. 9

’

NO. 12

’

11728–11736

’

ARTICLE

18. Jin, H. J.; Lee, S. H.; Kim, T. H.; Park, J.; Song, H. S.; Park, T. H.; Hong, S. Nanovesicle-Based Bioelectronic Nose Platform Mimicking Human Olfactory Signal Transduction. Biosens. Bioelectron. 2012, 35, 335–341. 19. Song, H. S.; Jin, H. J.; Ahn, S. R.; Kim, D.; Lee, S. H.; Kim, U. K.; Simons, C. T.; Hong, S.; Park, T. H. Bioelectronic Tongue Using Heterodimeric Human Taste Receptor for the Discrimination of Sweeteners with Human-like Performance. ACS Nano 2014, 8, 9781–9789. 20. Lee, M.; Im, J.; Lee, B. Y.; Myung, S.; Kang, J.; Huang, L.; Kwon, Y. K.; Hong, S. Linker-Free Directed Assembly of High-Performance Integrated Devices Based on Nanotubes and Nanowires. Nat. Nanotechnol. 2006, 1, 66–71. 21. Park, J.; Lim, J. H.; Jin, H. J.; Namgung, S.; Lee, S. H.; Park, T. H.; Hong, S. A Bioelectronic Sensor Based on Canine Olfactory Nanovesicle-Carbon Nanotube Hybrid Structures for the Fast Assessment of Food Quality. Analyst 2012, 137, 3249–3254. 22. Kim, T. H.; Lee, S. H.; Lee, J.; Song, H. S.; Oh, E. H.; Park, T. H.; Hong, S. Single-Carbon-Atomic-Resolution Detection of Odorant Molecules Using a Human Olfactory ReceptorBased Bioelectronic Nose. Adv. Mater. 2009, 21, 91–94. 23. Kim, B.; Lee, J.; Namgung, S.; Kim, J.; Park, J. Y.; Lee, M. S.; Hong, S. DNA Sensors Based on CNT-FET with Floating Electrodes. Sens. Actuators, B 2012, 169, 182–187. 24. Dong, H.; Dunn, J.; Lytton, J. Stoichiometry of the Cardiac Naþ/Ca2þ Exchanger NCX1.1 Measured in Transfected HEK Cells. Biophys. J. 2002, 82, 1943–1952. 25. Jin, H. J.; An, J. M.; Park, J.; Moon, S. J.; Hong, S. 00 ChemicalPain Sensor00 Based on Nanovesicle-Carbon Nanotube Hybrid Structures. Biosens. Bioelectron. 2013, 49, 86–91. 26. Kim, T. H.; Song, H. S.; Jin, H. J.; Lee, S. H.; Namgung, S.; Kim, U. K.; Park, T. H.; Hong, S. 00 Bioelectronic Super-Taster00 Device Based on Taste Receptor-Carbon Nanotube Hybrid Structures. Lab Chip 2011, 11, 2262–2267. 27. Van Boven, M.; Daenens, P.; Cokelaere, M.; Decuypere, E. Extraction and Liquid Chromatographic Method for the Determination of Simmondsin in Plasma. J. Chromatogr., Biomed. Appl. 1994, 655, 281–285. 28. Yoshii, K.; Yokouchi, C.; Kurihara, K. Synergistic Effects of 50 -Nucleotides on Rat Taste Responses to Various AminoAcids. Brain Res. 1986, 367, 45–51. 29. Li, X. D.; Staszewski, L.; Xu, H.; Durick, K.; Zoller, M.; Adler, E. Human Receptors for Sweet and Umami Taste. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 4692–4696. 30. Fuke, S.; Konosu, S. Taste-Active Components in Some Foods - A Review of Japanese Research. Physiol. Behav. 1991, 49, 863–868. 31. Drake, S. L.; Whetstine, M. E. C.; Drake, M. A.; Courtney, P.; Fligner, K.; Jenkins, J.; Pruitt, C. Sources of Umami Taste in Cheddar and Swiss Cheeses. J. Food Sci. 2007, 72, S360– S366. 32. Ninomiya, K. Natural Occurrence. Food Rev. Int. 1998, 14, 177–211. 33. Yamaguchi, S. The Synergistic Taste Effect of Monosodium Glutamate and Disodium 50 -Inosinate. J. Food Sci. 1967, 32, 473–478. 34. Lopez Cascales, J. J.; Costa, S. D. O.; de Groot, B. L.; Walters, D. E. Binding of Glutamate to the Umami Receptor. Biophys. Chem. 2010, 152, 139–144. 35. Pick, H.; Schmid, E. L.; Tairi, A. P.; Ilegems, E.; Hovius, R.; Vogel, H. Investigating Cellular Signaling Reactions in Single Attoliter Vesicles. J. Am. Chem. Soc. 2005, 127, 2908–2912. 36. Christensen, S. M.; Stamou, D. G. Sensing-Applications of Surface-Based Single Vesicle Arrays. Sensors 2010, 10, 11352–11368. 37. Jung, J. W.; Kim, J. H.; Pfeiffer, R.; Ahn, Y. J.; Page, T. L.; Kwon, H. W. Neuromodulation of Olfactory Sensitivity in the Peripheral Olfactory Organs of the American Cockroach, Periplaneta Americana. PLoS One 2013, 8, e81361. 38. Traut, R. R.; Bollen, A.; Sun, T. T.; Hershey, J. W. B.; Sundberg, J.; Pierce, L. R. Methyl 4-Mercaptobutyrimidate as a Cleavable Crosslinking Reagent and Its Application to Escherichia-Coli 30s Ribosome. Biochemistry 1973, 12, 3266–3273.

11736

2015 www.acsnano.org