The identification and quantification of protein biomarkers are crucial for advancing precision and personalized medicine. Many unexplored proteins may play significant roles in various pathological conditions, particularly in hematological malignancies and solid tumors, where changes in protein expression are frequently observed. Current protein detection techniques in biofluids often struggle with nonspecific binding, leading to high background noise and a limited dynamic range, especially at low analyte concentrations. Therefore, a significant need exists for highly sensitive and specific protein-sensing methods with rapid signal response. Single-molecule sensing using nanopores and the resistive-pulse technique offers advantages such as high sensitivity and adaptability to parallel recording technologies. However, current nanopore approaches often require the protein to enter the pore interior, leading to steric hindrance and limitations in detecting large proteins. This study aims to overcome these limitations by developing a novel nanopore sensor design that performs single-molecule protein detection outside the nanopore's confines.

Previous research has extensively explored the use of nanopores for protein analytics. Nanopores fabricated from various organic and inorganic materials have demonstrated the ability to identify and quantify peptides at single-amino acid resolution and reveal critical protein features, such as shape, size, post-translational modifications, enzymatic activity, and mechanical unfolding. Furthermore, nanopores have been employed to fingerprint proteins using enzymatic degradation. While offering many advantages, these methods typically require the protein to enter the pore, imposing steric constraints and hindering detection of larger proteins. An alternative is to detect proteins outside the nanopore using an external protein binder covalently attached to the nanopore. However, this requires a transducing mechanism and involves a complex optimization process for each protein analyte, making it less generalizable. Earlier studies have suggested limitations in detecting larger proteins with such external binders, likely due to steric constraints and limited interface accessibility. This study addresses these limitations by using a programmable antibody-mimetic binder attached to the nanopore's exterior.

The researchers developed a class of sensing elements consisting of an antibody-mimetic protein binder (monobody) engineered onto the tFhuA nanopore, a monomeric β-barrel scaffold, via a flexible tether. Monobodies were chosen for their smaller hydrodynamic radius, lack of disulfide bridges, and ease of expression and purification, allowing for straightforward adaptation to different target proteins. Three different monobodies (FN3SUMO, Mb4, and Adnectin1) targeting human small ubiquitin-related modifier 1 (hSUMO1), WD40 repeat protein 5 (WDR5), and the ectodomain of epidermal growth factor receptor (EGFR), respectively, were used to construct the sensors. AlphaFold2 was used for computational prediction of the three-dimensional structures of the fusion proteins. The sensors were refolded in detergents and reconstituted into a synthetic lipid bilayer at a single-molecule level. Single-channel electrical recordings were performed at a transmembrane potential of +40 mV (or +20 mV for Adnectin1-tFhuA) to monitor the interactions between the sensors and their target proteins. The presence of target proteins in the *cis* compartment produced characteristic current blockades whose amplitudes and durations were analyzed to determine association and dissociation rate constants, and equilibrium dissociation constants (K_D). Biolayer interferometry (BLI) was used to validate the fast kinetics of WDR5-Mb4 interactions. Steady-state fluorescence polarization (FP) anisotropy assays were conducted to determine the binding affinities of the sensors with their cognate analytes in solution. The sensors’ performance in a complex biofluid was assessed by detecting EGFR in the presence of 5% (v/v) fetal bovine serum (FBS).

The researchers successfully developed and validated three monobody-based nanopore sensors for the detection of hSUMO1, WDR5, and EGFR. The sensors exhibited high sensitivity and specificity, enabling label-free, real-time detection of the target proteins at single-molecule precision. Each target protein produced unique electrical signatures characterized by distinct current blockade amplitudes and durations. For hSUMO1 and WDR5, a single-exponential distribution was observed for both the release and capture durations. The frequency of capture events was linearly proportional to the analyte concentration, indicating a bimolecular association process, while the capture duration was independent of concentration, suggesting a unimolecular dissociation mechanism. Kinetic parameters (k_on, k_off, K_D) were obtained for both hSUMO1-FN3SUMO and WDR5-Mb4 interactions. Interestingly, the EGFR-Adnectin1 interaction displayed a bimodal protein recognition, resulting in two distinct populations of binding events (short-lived and long-lived) with different current blockade amplitudes and durations. Both short-lived and long-lived capture event frequencies were linearly dependent on the EGFR concentration. The detection of EGFR was also achieved successfully in the presence of 5% (v/v) FBS, demonstrating the sensor's robustness in complex biological samples. The EGFR concentration in the FBS-containing sample was successfully estimated using kinetic parameters obtained from measurements in buffer. Steady-state fluorescence polarization measurements confirmed the binding affinity of the sensors for hSUMO1 and WDR5, showing good agreement with the single-channel electrical recordings. The K_D values for the long-lived EGFR-captured events obtained using Adnectin1-tFhuA were consistent with previously reported values.

This study successfully demonstrates a highly generalizable approach for creating nanopore sensors for protein detection. The modular design, using a readily adaptable antibody-mimetic binder, allows for the creation of sensors tailored to various proteins, significantly expanding the application of nanopore technology. The ability to detect proteins outside the nanopore interior overcomes limitations associated with steric hindrance. The high sensitivity and specificity, along with the label-free and real-time nature of the detection, offer significant advantages over existing methods. The observation of bimodal protein recognition of EGFR highlights the capacity of this approach to reveal complex binding mechanisms. The successful detection of EGFR in the presence of FBS further validates the sensors' potential for real-world applications in complex biological samples. This technology holds promise for various applications, particularly in disease diagnostics and biomarker discovery.

This work presents a significant advancement in nanopore-based protein sensing. The modular, single-polypeptide nanopore sensors with adaptable antibody-mimetic binders offer a versatile and sensitive approach for detecting diverse proteins at single-molecule precision. The successful detection in complex media suggests great potential for application in biomedical diagnostics. Future research could focus on expanding the library of available monobodies to target a wider range of proteins, exploring the use of alternative protein binders, and integrating the sensors with high-throughput technologies for clinical diagnostics.

While this study demonstrates the high sensitivity and specificity of the nanopore sensors, further research is needed to fully characterize their performance across a broader range of proteins. Potential limitations include the need for specialized equipment for single-channel electrical recordings, the requirement for protein refolding, and potential interferences from other biomolecules in complex samples. Although detection in FBS was successful, more thorough testing in various biofluids is needed. The study primarily focused on three model proteins; more comprehensive testing with a wider array of target proteins, particularly those with very high association rates, will be necessary for robust validation.