Electrochemical biosensors convert biological signals, such as affinity binding, the presence of redox-active substances or inhibitors, and changes in pH, into electrical currents. With biocomponents as recognition elements, biosensors commonly demonstrate exceptional sensitivity and specificity for biomolecules, chemicals, and more. Utilizing electrodes as transducers can simplify electronic integration, easing the development of miniature and portable electronic devices, which significantly enhance biosensor applications in point-of-care (POC) diagnostics. An exemplary POC biosensor leading the way is the glucometer which is specifically designed for diabetes patients to monitor blood glucose levels. Expanding glucometers to encompass a broader spectrum of diagnostic applications opens the door to developing many sensors for various biomarkers in a cost-effective manner. Since frequent monitoring and rapid testing are critical in preventing drug resistance or cancer recurrence, we reengineer the glucometer into a cancer therapeutic sensor that can detect 4-hydroxytamoxifen (4-HT), a metabolite of the drug, tamoxifen, widely used in treating hormone receptor-positive breast cancer. Moreover, our electrochemical sensor can harness power from blood glucose for its operation and signal amplification. In creating such a highly integrated biosensor, our work demonstrates a broad interdisciplinary approach, capitalizing on recent innovations in protein engineering, electrochemical sensing, and electrical engineering.

The traditional glucometer operates by utilizing redox enzymes to catalyze glucose oxidation, resulting in a current that correlates to the glucose concentration. Leveraging protein engineering, we aimed to create a redox modular protein  that could rapidly transmit the 4-HT signal via glucose oxidation. We selected pyrroloquinoline quinone -glucose dehydrogenase (PQQ-GDH) as the base redox enzyme for its high turnover rate in catalyzing glucose oxidation. We then introduced the estrogen receptor ligand-binding domain (ER-LBD) to this GDH. We anticipated that the 4-HT induced conformational changes in the ER-LBD would be transmitted through the fused protein matrix, thereby modulating the redox activity of the GDH (Figure 1). 

However, a looming problem arose: where should we insert this domain into GDH to construct such a redox modular protein? Our solution- Everywhere. We built a saturated insertion library in which the ER-LBD was located between every residue of the GDH. By screening this library for areas that can tolerate insertion disruptions, as well as areas that propagate 4-HT signal to regulate glucose oxidation, we discovered this GDH was amenable to domain insertion between over 50% of its residues, particularly at the outer layer of beta-sheet propellers (Figure 2). We then took this pool of variants and further screened them based on their response to 4-HT. Interestingly, a significant fraction of these 4-HT regulatable sites were found at the dimerization interface of GDH. This alluded to the idea that enzymatic activity was being modulated by manipulating secondary and quaternary structures.

An insertion variant GDH-5E, where the ER-LBD was inserted following residue Threonine 5 of the GDH, was purified due to its exceptional response during whole-cell screening. However, this favorable outcome did not persist in enzymatic assay. We assume this difference is due to the transition from the E. coli cytosol to the buffer medium. Since the modest effect in buffer conditions was not sufficient to generate a sensor, we decided to increase the effect of LBD on the GDH-5E. Our mapping analysis revealed that flexible loops and alpha-helix are essential structures for signal propagation in GDH. Consequently, we inserted a flexible linker into GDH-5E, creating GDH-5E+ with improved signal. GDH-5E+ serves as an effective modulator, exhibiting rapid responses across the range of 4-HT concentrations in serum. GDH-5E+ also responds to other cancer therapeutics like hexestrol, diethylstilbestrol, and lasoxifene with good selectivity over estradiol (ES), the primary female sex hormone. Therefore, we employed GDH-5E+ as the redox modulator to construct 4-HT electrochemical sensors.

In a manner analogous to glucometry, we transmitted the signal of 4-HT into an electrical current. This involved immobilizing the GDH-5E+ on an electrode with a redox polymer, thus facilitating electron flow from protein to an electrode. During tests with human blood samples, we observed oxidative current due to a persistent baseline concentration of glucose in the blood (average 5.6 mM in adults). 4-HT competes with ES for binding sites, consequently slowing down or halting tumor cell growth in breast tissue. Notably, blood samples with 4-HT showed a lower current response than the samples with ES within 4 min. Recognizing the potential effect baseline glucose fluctuations may have on 4-HT detection, we introduced an additional GDH-coated electrode. Subsequently, we normalized the current from GDH-5E+ to the current from GDH within a single test, effectively differentiating the 4-HT signal from that of blood glucose (Figure 3). 

In addition to transmitting signals, blood glucose could also serve as an energy source to power the sensor itself and its signal amplifier. To demonstrate this, we paired the GDH-5E+ electrode with a laccase (an enzyme that catalyzes oxygen reduction) electrode that reduces oxygen at higher potential. Consequently, electrons generated during glucose oxidation spontaneously flow through an external circuit to reduce oxygen, thus producing power. When 4-HT inhibited glucose oxidation, a decrease in both current and power occurred, conveying information about the presence of 4-HT. We also integrated this self-powered sensor with an organic electrochemical transistor, serving as a signal amplifier. This coupling resulted in a threefold signal enhancement while generating signals in the milliampere range. The capability to harvest energy from the blood sample has an incredible upside- it could remove the need for batteries in POC devices. We acknowledge that our current prototype does not allow for real-time or in-situ sensing as a wearable sensor, but could in the future by layering multiple fuel cells in series for example. Given its simplicity of operation, rapid signal generation, and high compatibility, we anticipate that our system will be synergistic with other up-and-coming technologies to create easy-to-use, self-powered POC devices.

Looking back at the evolutionary history of glucometers, we are deeply impressed by the transition from single electrochemical testing to today's continuous glucose monitoring devices. This remarkable progress is the result of successfully integrating technological expertise from various fields, a model we were inspired by in developing our biosensor. We foresee collaboration between synthetic biology, protein engineering, and wearable electronics opening the door for wireless, implanted sensors for a myriad of biomarkers. This could further lead to POC prevention, diagnosis, and personalized treatment at a small cost. As glucometers have benefited millions of people with diabetes, these adapted glucometers will continue to provide individual healthcare solutions for various needs.