- Email: [email protected]
Continuous operation of an ultra-low-power microcontroller using glucose as the sole energy source
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Continuous operation of an ultra-low-power microcontroller using glucose as the sole energy source Inyoung Leea, Takashi Sodeb, Noya Loewc, Wakako Tsugawaa,c,d, Christopher Robin Lowec,e, ⁎ Koji Sodea,b,c,d, a Department of Industrial Technology and Innovation, Graduate School of Engineering, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan b Ultizyme International Ltd., 1-13-16 Minami, Meguro, Tokyo 152-0013, Japan c Institute of Global Innovation Research, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan d Department of Biotechnology and Life Science, Graduate School of Engineering, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan e Institute of Biotechnology, University of Cambridge, Tennis Court Road, Cambridge CB2 1QT, UK
A R T I C L E I N F O
A BS T RAC T
Keywords: Microcontroller BioCapacitor Enzyme fuel cell Direct electron transfer Implantable artificial organs Self-powered Glucose sensor
An ultimate goal for those engaged in research to develop implantable medical devices is to develop mechatronic implantable artificial organs such as artificial pancreas. Such devices would comprise at least a sensor module, an actuator module, and a controller module. For the development of optimal mechatronic implantable artificial organs, these modules should be self-powered and autonomously operated. In this study, we aimed to develop a microcontroller using the BioCapacitor principle. A direct electron transfer type glucose dehydrogenase was immobilized onto mesoporous carbon, and then deposited on the surface of a miniaturized Au electrode (7 mm2) to prepare a miniaturized enzyme anode. The enzyme fuel cell was connected with a 100 μF capacitor and a power boost converter as a charge pump. The voltage of the enzyme fuel cell was increased in a stepwise manner by the charge pump from 330 mV to 3.1 V, and the generated electricity was charged into a 100 μF capacitor. The charge pump circuit was connected to an ultra-low-power microcontroller. Thus prepared BioCapacitor based circuit was able to operate an ultra-low-power microcontroller continuously, by running a program for 17 h that turned on an LED every 60 s. Our success in operating a microcontroller using glucose as the sole energy source indicated the probability of realizing implantable self-powered autonomously operated artificial organs, such as artificial pancreas.
1. Introduction An ultimate goal for those engaged in research to develop implantable medical devices is to develop mechatronic implantable artificial organs, such as an artificial pancreas. Such devices would comprise at least a sensor module, an actuator module, and a controller module (Fig. 1). The sensor module recognizes the marker molecules to diagnose the patient and/or to monitor pharmaceutical compounds being used as medication. The actuator module delivers pharmaceutical compounds for medication. The controller module combines the information from the sensor module with a programmed database to operate the actuator module to achieve the appropriate dose of pharmaceutical compounds for medication. One representative challenge in this field is the development of closed-loop artificial pancreas. Recent progress in continuous glucose monitoring systems and mobile
⁎
device technology allows us to operate an insulin pump based on the sensing signals using a mobile device as the controller unit (e.g., artificial pancreas @ home, https://ec.europa.eu/digital-singlemarket/en/news/artificial-pancreas-whats-status). However, these approaches are still awaiting further investigation to realize associated medical devices. An ideal mechatronic implantable artificial organ should be autonomously operated and be self-powered. To achieve this, substantial effort has been made on developing biofuel cell-based medical device operations, such as self-powered glucose sensors (Katz et al., 2001, Katz and Willner, 2003; Mano et al., 2004; Kakehi et al., 2007; Cinquin et al., 2010; Falk et al., 2013; Cadet et al., 2016). However, enzyme fuel cells have the inherent problem that the theoretical voltage of a singleenzyme fuel cell is low because the voltage of the enzyme fuel cell depends on the redox potential of mediators and cofactors. In addition,
Correspondence to: Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan. E-mail address: [email protected] (K. Sode).
http://dx.doi.org/10.1016/j.bios.2016.09.095 Received 16 June 2016; Received in revised form 25 September 2016; Accepted 26 September 2016 Available online xxxx 0956-5663/ © 2016 Elsevier B.V. All rights reserved.
Please cite this article as: Lee, I., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.09.095
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
I. Lee et al.
immobilized on a micro-electrode, a capacitor connected with a power boost converter, and an ultra-low-power microcontroller were combined, and the operation of microcontroller by glucose as the sole energy source was achieved. 2. Materials and methods 2.1. Materials In this study, bacterial FAD-dependent glucose dehydrogenase (FADGDH) was used, which is capable of direct electron transfer. FADGDH comprises three subunits, namely the catalytic subunit, the cytochrome c subunit, and the small subunit. A recombinant FADGDH complex was prepared using the expression vectors pTrc99A, containing the structural gene for the FADGDH complex, and pACYC184, containing the structural genes for cytochrome c maturation (pEC86); the vectors were transformed into Escherichia coli strain BL21 (DE3) and cultivated as previously described (Tsuya et al., 2006). Bilirubin oxidase (BOD) was kindly donated by Amano Enzyme Inc. (Aichi, Japan). Ketjen black, ECP600JD, was purchased from Mitsubishi Chemical Corporation (Tokyo, Japan). Mesoporous carbon (Cnovel) was donated by Toyo Tanso Co., Ltd. (Osaka, Japan). 1-Pyrenebutyric acid N-hydroxysuccinimide ester was purchased from Sigma-Aldrich (St. Louis, MO). Au rod electrode, Pt mesh, and Ag/AgCl reference electrode were purchased from BAS Inc. (Tokyo, Japan). Further, 25% (w/v) glutaraldehyde solution was purchased from Wako Pure Chemical Industries, Ltd. (Oosaka, Japan), and Triton X-100 was purchased from Kanto Chemical (Tokyo, Japan).
Fig. 1. Conceptual diagram of mechatronic implantable artificial organs. For the operation of self-powered sensor-based autonomous artificial organs, a controller module, i.e. a self-powered computer, is essential. The controller module combines the information from the sensor module with a programmed database to operate the actuator module to achieve the appropriate dose of pharmaceutical compounds for medication. An ideal mechatronic implantable artificial organ should be autonomously operated and be self-powered based on BioCapacitor principle.
the power of an enzyme fuel cell, which is small enough to be implantable, might not be sufficiently high due to the small electrode surface area. To increase voltage and power, enzyme fuel cells could be arranged in series and in parallel. However, these designs are difficult to apply in implantable sensing devices. These limitations inspired us to develop a novel principle called BioCapacitor (Hanashi et al., 2009). In BioCapacitor, an enzyme fuel cell is connected with a capacitor via a charge pump. The charge pump steps up the voltage, and the electricity generated from the biofuel cell is charged in a capacitor. A high voltage with sufficient temporary power can be generated by the BioCapacitor principle. This principle is currently applied for the construction of several biosensing systems using enzyme fuel cells, which can operate various devices. We have already reported that autonomous sensing devices and actuator devices can be operated using this simple and innovative principle (Hanashi et al., 2011, 2012, 2014). This means that among the three modules necessary for mechatronic implantable artificial organs, the sensor module and the actuator module can be self-powered, whereas a controller module is yet to be developed. One challenging field in which the “BioCapacitor” has yet to be exploited is the realization of the unmet technology of operation of a computer (microprocessor) using glucose as the sole energy source. In other words, the development of a self-powered computer by employing the BioCapacitor principle is needed, which is inevitable to develop programmable autonomous biodevices to operate self-powered sensor-based autonomous drug delivery systems (Sode et al., 2016). Therefore, in this study, we demonstrate a self-powered microcontroller using glucose as the sole energy source. A microcontroller is a small computer including a central processing unit (CPU), random access memory (RAM), read only memory (ROM), input/output (I/O), and a timer. In a microcontroller, the memory stores the instructions (program) in advance and the CPU executes these instructions. Subsequently, the CPU of the microcontroller processes input data in accordance with the instructions and enables the output of data in different forms. When power is supplied to the microcontroller, this series of processes is automatic. Therefore, if the microcontroller can be operated using glucose as the sole energy source, we can construct a self-powered, stand-alone, implantable, and autonomous biodevice operated by the program. To realize this idea, an enzyme fuel cell consisting of a direct electron transfer type glucose dehydrogenase
2.2. Electrode preparation For electrode preparation, 10 mM 1-pyrenebutyric acid N-hydroxysuccinimide ester solution was prepared using acetone as a solvent. Ketjen black carbon ink was prepared by mixing 15 mg of Ketjen black with 670 μl of ultrapure water and 30 μl of Triton X-100. The enzyme anode was prepared using FADGDH complex-immobilized mesoporous carbon, as follows. Fifty-six micrograms of mesoporous carbon particles were incubated with 10 mM 1-pyrenebutyric acid N-hydroxysuccinimide ester for 1 h. The mixed solution was dried for 1 h and then supplemented with 7.9 mg/ml FADGDH complex and 50 mM HEPES buffer (pH 8.0). Next, 2% sucrose was added to the mixed enzyme solution containing 1-pyrenebutyric acid N-hydroxysuccinimide ester and FADGDH complex. A total of 7 μl of each enzyme solution was dropped onto the miniaturized Au electrode (7 mm2). The electrode was cross-linked in 25% glutaraldehyde vapor for 1 h and washed with 10 mM Tris-HCl buffer (pH 7.0). The anode was then stored in 100 mM potassium phosphate buffer (PPB; pH 7.0, electrolyte composition; KH2PO4+K2HPO4) until use. The enzyme cathode was prepared using BOD-immobilized Ketjen black, as follows. Two hundred microliters of Ketjen black ink, 300 μl of 20 mg/ml BOD solution, and 500 μl of 100 mM PPB (pH 7.0) were mixed, and then 200 μl of this solution was dropped onto a Pt mesh (80 mesh). The electrode was cross-linked in 25% glutaraldehyde vapor for 1 h and washed with 10 mM Tris-HCl buffer (pH 7.0). The cathode was then stored in 100 mM PPB (pH 7.0) until use. 2.3. Electrochemical characterization of the enzyme anode Chronoamperometric evaluation of the enzyme anode (+0.4 V vs. Ag/AgCl) was conducted using a 10-ml water jacket cell. The constructed anode, Ag/AgCl, and Pt wire were used as a working electrode, a reference electrode, and a counter electrode, respectively. All electrochemical characterizations were performed at 37 °C in 100 mM PPB (pH 7.0). Chronoamperometric evaluation was carried out in triplicate, using independently prepared three enzyme anodes. 2
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
I. Lee et al.
complex was immobilized onto mesoporous carbon using 1-pyrenebutyric acid N-hydroxysuccinimide ester, as described previously (Tsujimura et al., 2014; Viswanathan et al., 2015). Fig. 2A shows the result of chronoamperometric evaluation. This result shows that the prepared miniaturized enzyme anode exhibited high current (120 μA) with high current density (1.7 mA/cm2) in the presence of 20 mM glucose. Fig. 2B shows the representative polarization curve of the enzyme fuel cell. The results of triplicated investigations are shown in Supplementary information S3. The dependence of power and current on the cell voltage obtained from triplicated investigation revealed that the open-circuit voltage was 640 mV and the maximum current density and maximum power density were 1.87 ± 0.15 mA/cm2 at 200 Ω, and 0.36 ± 0.01 mW/cm2 at 5 kΩ, respectively (Fig. 2B and Supplementary information S3).
2.4. Polarization curve measurement of the enzyme fuel cell To construct an enzyme fuel cell, each enzyme anode which was subjected to chronoamperometric evaluation in Section 2.3. was immersed in a 10-ml water jacket cell with a freshly prepared cathode. Here, 100 mM PPB (pH 7.0) was used as the reaction solution, including 20 mM glucose. Next, the enzyme fuel cell was connected to an external variable-load resistance (model 278620; Yokogawa Electric Corporation, Tokyo, Japan), and the voltage generated by the enzyme fuel cell was measured. Polarization curve was measured in triplicate, by monitoring current and voltage of cells by changing resistances, using independently prepared three enzyme fuel cells. 2.5. Operation of ultra-low-power microcontroller by BioCapacitor principle
3.2. Operation of ultra-low-power microcontroller by the BioCapacitor principle
To construct the enzyme fuel cell, the enzyme anode and cathode were immersed in a 10-ml water jacket cell. Here, 100 mM PPB (pH 7.0) was used as the electrolyte solution. Next, the enzyme fuel cell was connected with a 100 μF capacitor and a power boost converter (BQ25504; Texas Instruments, Dallas, USA) as a charge pump. The voltage of the enzyme fuel cell was increased in a stepwise manner by the charge pump from 330 mV to 3.1 V, and the generated electricity was charged into a 100 μF capacitor. The charge pump circuit was connected to the ultra-low-power microcontroller (MSP430FR5739; Texas Instruments). The electronic circuit of the microcontroller operating system and a photograph of the experimental set-up of the system are shown in Supplementary information S1 and S2, respectively. We created a program for the microcontroller as follows: (1) The microcontroller executes a cycle of normal operation every 60 s in the operating state (voltage supply: 2–3 V, operating frequency of the microcontroller: 8 MHz). (2) After operation of the microcontroller, it is switched to sleep mode for 60 s (voltage supply: 2–3 V, the microcontroller is in a stopped state). (3) By the flashing of an LED, which is installed in the microcontroller, maximum power consumption is maintained at a constant level during normal operation, and the operation of the program can be visually confirmed.
Next, an ultra-low-power microcontroller was operated by combining with an enzyme fuel cell connected with a capacitor and a power booster to form a BioCapacitor. The operation of microcontroller was monitored by running the installed program. In this program, when the microcontroller is powered up, the CPU instructs that the LED should be switched on and lit up using charged power from the capacitor. Subsequently, the microcontroller maintains the LED in a lit state for a fixed time. After this fixed time, the microcontroller instructs that the LED should be switched off. Then, its CPU of microcontroller instructs the entry into sleep mode, during which only approximately 0.7 μA current is used. When the microcontroller enters sleep mode, energy in the capacitor is recharged. After 60 s, the microcontroller repeats the program of the LED switching on and off according to the program (the algorithm of this program is shown in Supplementary information S4). Therefore, the operation of the microcontroller can be confirmed by observing the voltage of the capacitor because the microcontroller operates using only charged power from the capacitor. Fig. 3 shows the time courses of the continuous operation of the microcontroller, according to the installed program. The operation of the microcontroller was observed by monitoring the voltage of the capacitor. The charged power from the capacitor was used to operate the microcontroller program. According to the program instructions, the microcontroller lit up the LED using charged power from the capacitor; therefore, the voltage of the capacitor decreased from 3 to 2.5 V (Fig. 3A). The microcontroller maintained the LED in a lit state for 125 μs, after which it instructed the LED to be switched off. After
3. Results 3.1. Electrochemical characterization of electrode To achieve high current density of the enzyme anode, an FADGDH
Fig. 2. Electrochemical characterization. (A) Chronoamperometric evaluation of the FADGDH immobilized anode using mesoporous carbon and 1-pyrenebutyric acid Nhydroxysuccinimide ester at +0.4 V (vs. Ag/AgCl) in 100 mM potassium phosphate buffer (pH 7.0). Circles indicate mean current values from the evaluations carried out in triplicate, and error bars indicate the standard deviation (N=3). (B) Representative polarization curve of an enzyme fuel cell based on direct electron transfer with an FADGDH immobilized anode and BOD immobilized cathode. Reaction solution was 100 mM potassium phosphate buffer (pH 7.0) containing 20 mM glucose. Filled squares show the dependence of current density on cell voltage and open squares show the dependence of power density on cell voltage.
3
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
I. Lee et al.
Fig. 3. Time course of the continuous operation of the microcontroller using the BioCapacitor principle, according to the installed program, (A) The microcontroller lit up the LED using charged power from the capacitor. (B) When the microcontroller entered sleep mode, energy in the capacitor was recharged. (Charging time: 8.3 s) (C) Every 60 s, the microcontroller woke from the sleep mode and operated the program repeatedly. (D) The microcontroller program worked normally for 17 h using only glucose.
density, which consequently realized the continuous operation of the microcontroller, even though a miniaturized anode suitable for implantation was used. Although further investigations are required, our success in operating a microcontroller using glucose as the sole energy source indicated the possibility of realizing implantable self-powered autonomously operated drug delivery systems, or artificial organs, such as a mechatronic artificial pancreas.
turning off the LED, the microcontroller entered sleep mode, drawing only approximately 0.7 μA current. When the microcontroller entered sleep mode, energy in the capacitor was recharged (from 2.5 to 3 V), which took just 8.3 s (Fig. 3B). Every 60 s, the microcontroller woke from the sleep mode and repeated to operate the program (Fig. 3C). The microcontroller continuously turned the LED on and off, and the microcontroller program worked for 17 h using only glucose (Fig. 3D). The electric power required for the program of the microcontroller was 1.35 mW. These results demonstrated that BioCapacitor equipped with a miniaturized enzyme anode was capable of continuously operating an ultra-low-power microcontroller by running a program for 17 h that turned on an LED every 60 s.
5. Conclusion In this study, we have demonstrated that an ultra-low-power microcontroller can be continuously operated using glucose as the sole energy source, by employing the BioCapacitor principle. A program installed in an ultra-low-power microcontroller was run for 17 h. Our success indicated the probability of realizing implantable self-powered autonomously operated artificial organs, such as a mechatronic artificial pancreas.
4. Discussion In this study, we have demonstrated that an ultra-low-power microcontroller can be continuously operated using glucose as the sole energy source, by employing the BioCapacitor principle. Because several programs can be installed on such a microcontroller, which is necessary to operate biodevices, this achievement suggests the probability of a future development of a self-powered controller module. MacVittie et al. (2015) previously reported on the successful operation of a microcontroller using only the power from an enzyme fuel cell inserted in an orange. The size of an orange allowed the use of a large electrode (7.5 cm2), which could generate sufficient power in combination with a supercapacitor to operate the microcontroller. We employed the most ultra-low-power microcontroller that is currently available, which required 10 times less current (0.7 μA) during its sleep mode than that was reported by MacVittie et al. (2015). Moreover, the employment of bacterium-derived FADGDH capable of direct electron transfer immobilized on mesoporous carbon using 1-pyrenebutyric acid N-hydroxysuccinimide ester achieved a high anodic current
Acknowledgements The authors are grateful to Dr. Hirotaka Sato, School of Mechanical and Aerospace Engineering, Nanyang Technological University, Singapore, for introducing the power boost converter and for helpful discussions. The authors are also grateful to Dr. Seiya Tsujimura, Faculty of Pure and Applied Science, Tsukuba University, Japan, for the helpful discussion in revising this manuscript. We would like to thank Amano Enzyme Inc. (Aichi, Japan) for supplying bilirubin oxidase, and Toyo Tanso Co. Ltd. (Tokyo Japan) for providing mesoporous carbon. The authors would like also to thank Crimson Interactive Pvt. Ltd. (Mumbai, India) for the English language review service. 4
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
I. Lee et al.
Hanashi, T., Yamazaki, T., Tsugawa, W., Ikebukuro, K., Sode, K., 2012. Electrochemistry 80, 367–370. Hanashi, T., Yamazaki, T., Tanaka, H., Ikebukuro, K., Tsugawa, W., Sode, K., 2014. Sens. Actuators B: Chem. 196, 429–433. Kakehi, N., Yamazaki, T., Tsugawa, W., Sode, K., 2007. Biosens. Bioelectron. 22, 2250–2255. Katz, E., Buckmann, A.F., Willner, I., 2001. J. Am. Chem. Soc. 123, 10752–10753. Katz, E., Willner, I., 2003. J. Am. Chem. Soc. 125, 6803–6813. MacVittie, K., Conlon, T., Katz, E., 2015. Bioelectrochemistry 106, 28–33. Mano, N., Mao, F., Heller, A., 2004. Chembiochem 5, 1703–1705. Sode, K., Yamazaki, T., Lee, I., Hanashi, T., Tsugawa, W., 2016. Biosens. Bioelectron. 76, 20–28. Tsuya, T., Ferri, S., Fujikawa, M., Yamaoka, H., Sode, K., 2006. J. Biotechnol. 123, 127–136. Tsujimura, S., Murata, K., Akatsuka, W., 2014. J. Am. Chem. Soc. 136, 14432–14437. Viswanathan, S., Narayanan, T.N., Aran, K., Fink, K.D., Paredes, J., Ajayan, P.M., Filipek, S., Miszta, P., Tekin, H.C., Inci, F., Demirci, U., Li, P., Bolotin, K.I., Liepmann, D., Renugopalakrishanan, V., 2015. Mater. Today 18, 513–522.
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.bios.2016.09.095. References Cadet, M., Gounel, S., Stines-Chaumeil, C., Brilland, X., Rouhana, J., Louerat, F., Mano, N., 2016. Biosens. Bioelectron. 83, 60–67. Cinquin, P., Gondran, C., Giroud, F., Mazabrard, S., Pellissier, A., Boucher, F., Alcaraz, J., Gorgy, K., Lenouvel, F., Mathe, S., Porcu, P., Cosnier, S., 2010. PLoS ONE 5, e10476. Falk, M., Andoralov, V., Silow, M., Toscano, M.D., Shleev, S., 2013. Anal. Chem. 85, 6342–6348. Hanashi, T., Yamazaki, T., Tsugawa, W., Ferri, S., Nakayama, D., Tomiyama, M., Ikebukuro, K., Sode, K., 2009. Biosens. Bioelectron. 24, 1837–1842. Hanashi, T., Yamazaki, T., Tsugawa, W., Ikebukuro, K., Sode, K., 2011. J. Diabetes Sci. Technol. 5, 1030–1035.
5
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Continuous operation of an ultra-low-power microcontroller using glucose as the sole energy source Inyoung Leea, Takashi Sodeb, Noya Loewc, Wakako Tsugawaa,c,d, Christopher Robin Lowec,e, ⁎ Koji Sodea,b,c,d, a Department of Industrial Technology and Innovation, Graduate School of Engineering, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan b Ultizyme International Ltd., 1-13-16 Minami, Meguro, Tokyo 152-0013, Japan c Institute of Global Innovation Research, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan d Department of Biotechnology and Life Science, Graduate School of Engineering, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan e Institute of Biotechnology, University of Cambridge, Tennis Court Road, Cambridge CB2 1QT, UK
A R T I C L E I N F O
A BS T RAC T
Keywords: Microcontroller BioCapacitor Enzyme fuel cell Direct electron transfer Implantable artificial organs Self-powered Glucose sensor
An ultimate goal for those engaged in research to develop implantable medical devices is to develop mechatronic implantable artificial organs such as artificial pancreas. Such devices would comprise at least a sensor module, an actuator module, and a controller module. For the development of optimal mechatronic implantable artificial organs, these modules should be self-powered and autonomously operated. In this study, we aimed to develop a microcontroller using the BioCapacitor principle. A direct electron transfer type glucose dehydrogenase was immobilized onto mesoporous carbon, and then deposited on the surface of a miniaturized Au electrode (7 mm2) to prepare a miniaturized enzyme anode. The enzyme fuel cell was connected with a 100 μF capacitor and a power boost converter as a charge pump. The voltage of the enzyme fuel cell was increased in a stepwise manner by the charge pump from 330 mV to 3.1 V, and the generated electricity was charged into a 100 μF capacitor. The charge pump circuit was connected to an ultra-low-power microcontroller. Thus prepared BioCapacitor based circuit was able to operate an ultra-low-power microcontroller continuously, by running a program for 17 h that turned on an LED every 60 s. Our success in operating a microcontroller using glucose as the sole energy source indicated the probability of realizing implantable self-powered autonomously operated artificial organs, such as artificial pancreas.
1. Introduction An ultimate goal for those engaged in research to develop implantable medical devices is to develop mechatronic implantable artificial organs, such as an artificial pancreas. Such devices would comprise at least a sensor module, an actuator module, and a controller module (Fig. 1). The sensor module recognizes the marker molecules to diagnose the patient and/or to monitor pharmaceutical compounds being used as medication. The actuator module delivers pharmaceutical compounds for medication. The controller module combines the information from the sensor module with a programmed database to operate the actuator module to achieve the appropriate dose of pharmaceutical compounds for medication. One representative challenge in this field is the development of closed-loop artificial pancreas. Recent progress in continuous glucose monitoring systems and mobile
⁎
device technology allows us to operate an insulin pump based on the sensing signals using a mobile device as the controller unit (e.g., artificial pancreas @ home, https://ec.europa.eu/digital-singlemarket/en/news/artificial-pancreas-whats-status). However, these approaches are still awaiting further investigation to realize associated medical devices. An ideal mechatronic implantable artificial organ should be autonomously operated and be self-powered. To achieve this, substantial effort has been made on developing biofuel cell-based medical device operations, such as self-powered glucose sensors (Katz et al., 2001, Katz and Willner, 2003; Mano et al., 2004; Kakehi et al., 2007; Cinquin et al., 2010; Falk et al., 2013; Cadet et al., 2016). However, enzyme fuel cells have the inherent problem that the theoretical voltage of a singleenzyme fuel cell is low because the voltage of the enzyme fuel cell depends on the redox potential of mediators and cofactors. In addition,
Correspondence to: Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan. E-mail address: [email protected] (K. Sode).
http://dx.doi.org/10.1016/j.bios.2016.09.095 Received 16 June 2016; Received in revised form 25 September 2016; Accepted 26 September 2016 Available online xxxx 0956-5663/ © 2016 Elsevier B.V. All rights reserved.
Please cite this article as: Lee, I., Biosensors and Bioelectronics (2016), http://dx.doi.org/10.1016/j.bios.2016.09.095
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
I. Lee et al.
immobilized on a micro-electrode, a capacitor connected with a power boost converter, and an ultra-low-power microcontroller were combined, and the operation of microcontroller by glucose as the sole energy source was achieved. 2. Materials and methods 2.1. Materials In this study, bacterial FAD-dependent glucose dehydrogenase (FADGDH) was used, which is capable of direct electron transfer. FADGDH comprises three subunits, namely the catalytic subunit, the cytochrome c subunit, and the small subunit. A recombinant FADGDH complex was prepared using the expression vectors pTrc99A, containing the structural gene for the FADGDH complex, and pACYC184, containing the structural genes for cytochrome c maturation (pEC86); the vectors were transformed into Escherichia coli strain BL21 (DE3) and cultivated as previously described (Tsuya et al., 2006). Bilirubin oxidase (BOD) was kindly donated by Amano Enzyme Inc. (Aichi, Japan). Ketjen black, ECP600JD, was purchased from Mitsubishi Chemical Corporation (Tokyo, Japan). Mesoporous carbon (Cnovel) was donated by Toyo Tanso Co., Ltd. (Osaka, Japan). 1-Pyrenebutyric acid N-hydroxysuccinimide ester was purchased from Sigma-Aldrich (St. Louis, MO). Au rod electrode, Pt mesh, and Ag/AgCl reference electrode were purchased from BAS Inc. (Tokyo, Japan). Further, 25% (w/v) glutaraldehyde solution was purchased from Wako Pure Chemical Industries, Ltd. (Oosaka, Japan), and Triton X-100 was purchased from Kanto Chemical (Tokyo, Japan).
Fig. 1. Conceptual diagram of mechatronic implantable artificial organs. For the operation of self-powered sensor-based autonomous artificial organs, a controller module, i.e. a self-powered computer, is essential. The controller module combines the information from the sensor module with a programmed database to operate the actuator module to achieve the appropriate dose of pharmaceutical compounds for medication. An ideal mechatronic implantable artificial organ should be autonomously operated and be self-powered based on BioCapacitor principle.
the power of an enzyme fuel cell, which is small enough to be implantable, might not be sufficiently high due to the small electrode surface area. To increase voltage and power, enzyme fuel cells could be arranged in series and in parallel. However, these designs are difficult to apply in implantable sensing devices. These limitations inspired us to develop a novel principle called BioCapacitor (Hanashi et al., 2009). In BioCapacitor, an enzyme fuel cell is connected with a capacitor via a charge pump. The charge pump steps up the voltage, and the electricity generated from the biofuel cell is charged in a capacitor. A high voltage with sufficient temporary power can be generated by the BioCapacitor principle. This principle is currently applied for the construction of several biosensing systems using enzyme fuel cells, which can operate various devices. We have already reported that autonomous sensing devices and actuator devices can be operated using this simple and innovative principle (Hanashi et al., 2011, 2012, 2014). This means that among the three modules necessary for mechatronic implantable artificial organs, the sensor module and the actuator module can be self-powered, whereas a controller module is yet to be developed. One challenging field in which the “BioCapacitor” has yet to be exploited is the realization of the unmet technology of operation of a computer (microprocessor) using glucose as the sole energy source. In other words, the development of a self-powered computer by employing the BioCapacitor principle is needed, which is inevitable to develop programmable autonomous biodevices to operate self-powered sensor-based autonomous drug delivery systems (Sode et al., 2016). Therefore, in this study, we demonstrate a self-powered microcontroller using glucose as the sole energy source. A microcontroller is a small computer including a central processing unit (CPU), random access memory (RAM), read only memory (ROM), input/output (I/O), and a timer. In a microcontroller, the memory stores the instructions (program) in advance and the CPU executes these instructions. Subsequently, the CPU of the microcontroller processes input data in accordance with the instructions and enables the output of data in different forms. When power is supplied to the microcontroller, this series of processes is automatic. Therefore, if the microcontroller can be operated using glucose as the sole energy source, we can construct a self-powered, stand-alone, implantable, and autonomous biodevice operated by the program. To realize this idea, an enzyme fuel cell consisting of a direct electron transfer type glucose dehydrogenase
2.2. Electrode preparation For electrode preparation, 10 mM 1-pyrenebutyric acid N-hydroxysuccinimide ester solution was prepared using acetone as a solvent. Ketjen black carbon ink was prepared by mixing 15 mg of Ketjen black with 670 μl of ultrapure water and 30 μl of Triton X-100. The enzyme anode was prepared using FADGDH complex-immobilized mesoporous carbon, as follows. Fifty-six micrograms of mesoporous carbon particles were incubated with 10 mM 1-pyrenebutyric acid N-hydroxysuccinimide ester for 1 h. The mixed solution was dried for 1 h and then supplemented with 7.9 mg/ml FADGDH complex and 50 mM HEPES buffer (pH 8.0). Next, 2% sucrose was added to the mixed enzyme solution containing 1-pyrenebutyric acid N-hydroxysuccinimide ester and FADGDH complex. A total of 7 μl of each enzyme solution was dropped onto the miniaturized Au electrode (7 mm2). The electrode was cross-linked in 25% glutaraldehyde vapor for 1 h and washed with 10 mM Tris-HCl buffer (pH 7.0). The anode was then stored in 100 mM potassium phosphate buffer (PPB; pH 7.0, electrolyte composition; KH2PO4+K2HPO4) until use. The enzyme cathode was prepared using BOD-immobilized Ketjen black, as follows. Two hundred microliters of Ketjen black ink, 300 μl of 20 mg/ml BOD solution, and 500 μl of 100 mM PPB (pH 7.0) were mixed, and then 200 μl of this solution was dropped onto a Pt mesh (80 mesh). The electrode was cross-linked in 25% glutaraldehyde vapor for 1 h and washed with 10 mM Tris-HCl buffer (pH 7.0). The cathode was then stored in 100 mM PPB (pH 7.0) until use. 2.3. Electrochemical characterization of the enzyme anode Chronoamperometric evaluation of the enzyme anode (+0.4 V vs. Ag/AgCl) was conducted using a 10-ml water jacket cell. The constructed anode, Ag/AgCl, and Pt wire were used as a working electrode, a reference electrode, and a counter electrode, respectively. All electrochemical characterizations were performed at 37 °C in 100 mM PPB (pH 7.0). Chronoamperometric evaluation was carried out in triplicate, using independently prepared three enzyme anodes. 2
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
I. Lee et al.
complex was immobilized onto mesoporous carbon using 1-pyrenebutyric acid N-hydroxysuccinimide ester, as described previously (Tsujimura et al., 2014; Viswanathan et al., 2015). Fig. 2A shows the result of chronoamperometric evaluation. This result shows that the prepared miniaturized enzyme anode exhibited high current (120 μA) with high current density (1.7 mA/cm2) in the presence of 20 mM glucose. Fig. 2B shows the representative polarization curve of the enzyme fuel cell. The results of triplicated investigations are shown in Supplementary information S3. The dependence of power and current on the cell voltage obtained from triplicated investigation revealed that the open-circuit voltage was 640 mV and the maximum current density and maximum power density were 1.87 ± 0.15 mA/cm2 at 200 Ω, and 0.36 ± 0.01 mW/cm2 at 5 kΩ, respectively (Fig. 2B and Supplementary information S3).
2.4. Polarization curve measurement of the enzyme fuel cell To construct an enzyme fuel cell, each enzyme anode which was subjected to chronoamperometric evaluation in Section 2.3. was immersed in a 10-ml water jacket cell with a freshly prepared cathode. Here, 100 mM PPB (pH 7.0) was used as the reaction solution, including 20 mM glucose. Next, the enzyme fuel cell was connected to an external variable-load resistance (model 278620; Yokogawa Electric Corporation, Tokyo, Japan), and the voltage generated by the enzyme fuel cell was measured. Polarization curve was measured in triplicate, by monitoring current and voltage of cells by changing resistances, using independently prepared three enzyme fuel cells. 2.5. Operation of ultra-low-power microcontroller by BioCapacitor principle
3.2. Operation of ultra-low-power microcontroller by the BioCapacitor principle
To construct the enzyme fuel cell, the enzyme anode and cathode were immersed in a 10-ml water jacket cell. Here, 100 mM PPB (pH 7.0) was used as the electrolyte solution. Next, the enzyme fuel cell was connected with a 100 μF capacitor and a power boost converter (BQ25504; Texas Instruments, Dallas, USA) as a charge pump. The voltage of the enzyme fuel cell was increased in a stepwise manner by the charge pump from 330 mV to 3.1 V, and the generated electricity was charged into a 100 μF capacitor. The charge pump circuit was connected to the ultra-low-power microcontroller (MSP430FR5739; Texas Instruments). The electronic circuit of the microcontroller operating system and a photograph of the experimental set-up of the system are shown in Supplementary information S1 and S2, respectively. We created a program for the microcontroller as follows: (1) The microcontroller executes a cycle of normal operation every 60 s in the operating state (voltage supply: 2–3 V, operating frequency of the microcontroller: 8 MHz). (2) After operation of the microcontroller, it is switched to sleep mode for 60 s (voltage supply: 2–3 V, the microcontroller is in a stopped state). (3) By the flashing of an LED, which is installed in the microcontroller, maximum power consumption is maintained at a constant level during normal operation, and the operation of the program can be visually confirmed.
Next, an ultra-low-power microcontroller was operated by combining with an enzyme fuel cell connected with a capacitor and a power booster to form a BioCapacitor. The operation of microcontroller was monitored by running the installed program. In this program, when the microcontroller is powered up, the CPU instructs that the LED should be switched on and lit up using charged power from the capacitor. Subsequently, the microcontroller maintains the LED in a lit state for a fixed time. After this fixed time, the microcontroller instructs that the LED should be switched off. Then, its CPU of microcontroller instructs the entry into sleep mode, during which only approximately 0.7 μA current is used. When the microcontroller enters sleep mode, energy in the capacitor is recharged. After 60 s, the microcontroller repeats the program of the LED switching on and off according to the program (the algorithm of this program is shown in Supplementary information S4). Therefore, the operation of the microcontroller can be confirmed by observing the voltage of the capacitor because the microcontroller operates using only charged power from the capacitor. Fig. 3 shows the time courses of the continuous operation of the microcontroller, according to the installed program. The operation of the microcontroller was observed by monitoring the voltage of the capacitor. The charged power from the capacitor was used to operate the microcontroller program. According to the program instructions, the microcontroller lit up the LED using charged power from the capacitor; therefore, the voltage of the capacitor decreased from 3 to 2.5 V (Fig. 3A). The microcontroller maintained the LED in a lit state for 125 μs, after which it instructed the LED to be switched off. After
3. Results 3.1. Electrochemical characterization of electrode To achieve high current density of the enzyme anode, an FADGDH
Fig. 2. Electrochemical characterization. (A) Chronoamperometric evaluation of the FADGDH immobilized anode using mesoporous carbon and 1-pyrenebutyric acid Nhydroxysuccinimide ester at +0.4 V (vs. Ag/AgCl) in 100 mM potassium phosphate buffer (pH 7.0). Circles indicate mean current values from the evaluations carried out in triplicate, and error bars indicate the standard deviation (N=3). (B) Representative polarization curve of an enzyme fuel cell based on direct electron transfer with an FADGDH immobilized anode and BOD immobilized cathode. Reaction solution was 100 mM potassium phosphate buffer (pH 7.0) containing 20 mM glucose. Filled squares show the dependence of current density on cell voltage and open squares show the dependence of power density on cell voltage.
3
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
I. Lee et al.
Fig. 3. Time course of the continuous operation of the microcontroller using the BioCapacitor principle, according to the installed program, (A) The microcontroller lit up the LED using charged power from the capacitor. (B) When the microcontroller entered sleep mode, energy in the capacitor was recharged. (Charging time: 8.3 s) (C) Every 60 s, the microcontroller woke from the sleep mode and operated the program repeatedly. (D) The microcontroller program worked normally for 17 h using only glucose.
density, which consequently realized the continuous operation of the microcontroller, even though a miniaturized anode suitable for implantation was used. Although further investigations are required, our success in operating a microcontroller using glucose as the sole energy source indicated the possibility of realizing implantable self-powered autonomously operated drug delivery systems, or artificial organs, such as a mechatronic artificial pancreas.
turning off the LED, the microcontroller entered sleep mode, drawing only approximately 0.7 μA current. When the microcontroller entered sleep mode, energy in the capacitor was recharged (from 2.5 to 3 V), which took just 8.3 s (Fig. 3B). Every 60 s, the microcontroller woke from the sleep mode and repeated to operate the program (Fig. 3C). The microcontroller continuously turned the LED on and off, and the microcontroller program worked for 17 h using only glucose (Fig. 3D). The electric power required for the program of the microcontroller was 1.35 mW. These results demonstrated that BioCapacitor equipped with a miniaturized enzyme anode was capable of continuously operating an ultra-low-power microcontroller by running a program for 17 h that turned on an LED every 60 s.
5. Conclusion In this study, we have demonstrated that an ultra-low-power microcontroller can be continuously operated using glucose as the sole energy source, by employing the BioCapacitor principle. A program installed in an ultra-low-power microcontroller was run for 17 h. Our success indicated the probability of realizing implantable self-powered autonomously operated artificial organs, such as a mechatronic artificial pancreas.
4. Discussion In this study, we have demonstrated that an ultra-low-power microcontroller can be continuously operated using glucose as the sole energy source, by employing the BioCapacitor principle. Because several programs can be installed on such a microcontroller, which is necessary to operate biodevices, this achievement suggests the probability of a future development of a self-powered controller module. MacVittie et al. (2015) previously reported on the successful operation of a microcontroller using only the power from an enzyme fuel cell inserted in an orange. The size of an orange allowed the use of a large electrode (7.5 cm2), which could generate sufficient power in combination with a supercapacitor to operate the microcontroller. We employed the most ultra-low-power microcontroller that is currently available, which required 10 times less current (0.7 μA) during its sleep mode than that was reported by MacVittie et al. (2015). Moreover, the employment of bacterium-derived FADGDH capable of direct electron transfer immobilized on mesoporous carbon using 1-pyrenebutyric acid N-hydroxysuccinimide ester achieved a high anodic current
Acknowledgements The authors are grateful to Dr. Hirotaka Sato, School of Mechanical and Aerospace Engineering, Nanyang Technological University, Singapore, for introducing the power boost converter and for helpful discussions. The authors are also grateful to Dr. Seiya Tsujimura, Faculty of Pure and Applied Science, Tsukuba University, Japan, for the helpful discussion in revising this manuscript. We would like to thank Amano Enzyme Inc. (Aichi, Japan) for supplying bilirubin oxidase, and Toyo Tanso Co. Ltd. (Tokyo Japan) for providing mesoporous carbon. The authors would like also to thank Crimson Interactive Pvt. Ltd. (Mumbai, India) for the English language review service. 4
Biosensors and Bioelectronics xx (xxxx) xxxx–xxxx
I. Lee et al.
Hanashi, T., Yamazaki, T., Tsugawa, W., Ikebukuro, K., Sode, K., 2012. Electrochemistry 80, 367–370. Hanashi, T., Yamazaki, T., Tanaka, H., Ikebukuro, K., Tsugawa, W., Sode, K., 2014. Sens. Actuators B: Chem. 196, 429–433. Kakehi, N., Yamazaki, T., Tsugawa, W., Sode, K., 2007. Biosens. Bioelectron. 22, 2250–2255. Katz, E., Buckmann, A.F., Willner, I., 2001. J. Am. Chem. Soc. 123, 10752–10753. Katz, E., Willner, I., 2003. J. Am. Chem. Soc. 125, 6803–6813. MacVittie, K., Conlon, T., Katz, E., 2015. Bioelectrochemistry 106, 28–33. Mano, N., Mao, F., Heller, A., 2004. Chembiochem 5, 1703–1705. Sode, K., Yamazaki, T., Lee, I., Hanashi, T., Tsugawa, W., 2016. Biosens. Bioelectron. 76, 20–28. Tsuya, T., Ferri, S., Fujikawa, M., Yamaoka, H., Sode, K., 2006. J. Biotechnol. 123, 127–136. Tsujimura, S., Murata, K., Akatsuka, W., 2014. J. Am. Chem. Soc. 136, 14432–14437. Viswanathan, S., Narayanan, T.N., Aran, K., Fink, K.D., Paredes, J., Ajayan, P.M., Filipek, S., Miszta, P., Tekin, H.C., Inci, F., Demirci, U., Li, P., Bolotin, K.I., Liepmann, D., Renugopalakrishanan, V., 2015. Mater. Today 18, 513–522.
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.bios.2016.09.095. References Cadet, M., Gounel, S., Stines-Chaumeil, C., Brilland, X., Rouhana, J., Louerat, F., Mano, N., 2016. Biosens. Bioelectron. 83, 60–67. Cinquin, P., Gondran, C., Giroud, F., Mazabrard, S., Pellissier, A., Boucher, F., Alcaraz, J., Gorgy, K., Lenouvel, F., Mathe, S., Porcu, P., Cosnier, S., 2010. PLoS ONE 5, e10476. Falk, M., Andoralov, V., Silow, M., Toscano, M.D., Shleev, S., 2013. Anal. Chem. 85, 6342–6348. Hanashi, T., Yamazaki, T., Tsugawa, W., Ferri, S., Nakayama, D., Tomiyama, M., Ikebukuro, K., Sode, K., 2009. Biosens. Bioelectron. 24, 1837–1842. Hanashi, T., Yamazaki, T., Tsugawa, W., Ikebukuro, K., Sode, K., 2011. J. Diabetes Sci. Technol. 5, 1030–1035.
5