Molecular-electromechanical system for unamplified detection of trace analytes in biofluids

Detection of trace analytes or biomarkers in complex biofluids (down to 1–10 copies in 100 μl) is of great importance in fields such as biological research, precision medicine and early-stage diagnosis1,2,3,4. Until now, diverse sensing technologies have been studied; these include spectroscopy5,6, magnetic resonance imaging7,8, chromatography9,10, ion mobility spectrometry11,12, immunoassay13,14,15,16, PCR17,18, electrochemiluminescence19,20, surface-enhanced Raman spectroscopy21,22, surface plasmon resonance (SPR)23,24, and electrochemical sensors4,25, which have been developed for research and commercial applications. Although these technologies have achieved great progress, they suffer from insufficient sensitivity when detecting trace analytes in unamplified samples. To enhance the signal, analyte enrichment or amplification such as amplification of low amounts of nucleic acid by PCR is usually required26,27, which necessitates complicated preparation and increases testing time. More importantly, there is a trade-off between sensitivity and antifouling capability of preventing non-specific adsorption28. Non-specific adsorption of proteins, nucleic acids or other background molecules in biofluids may increase background noise, block receptor active sites and compromise the sensitivity of the assay. In addition, these technologies require costly instruments, usage of labels and well-trained operators and are not portable. These limitations are difficult to mitigate, partially because of a lack of strategy for actively manipulating sensing operations at the molecular level. In contrast, bio-recognition and specific response with remarkable precision happen in living systems. These sensing behaviors are dependent on precisely regulated molecular mechanisms such as protein synthesis29, metabolism30 and transmembrane signaling31, providing an alternative way to design artificial sensing systems operating in biofluids and overcome the aforementioned limitations.

Microelectromechanical systems (MEMSs) and nanoelectromechanical systems (NEMSs) integrate electrical and mechanical components at the microscale and nanoscale dimensions, respectively, to convert mechanical, chemical, biological or other sensing responses to electrical signals. From 196732, researchers started to combine the MEMS/NEMS technology with field-effect transistors (FETs). Recent studies show that by integration with FETs, sensitive MEMS/NEMS biosensors can be achieved because of efficient signal transduction resulting from external perturbation and the inherent capability of signal amplification of FETs33,34,35,36,37. Nevertheless, the precision is still far below that of living systems. Smaller dimensions of the sensing component mean higher sensitivity because a smaller sensing component undergoes a larger change in physical properties when subjected to external perturbations. Without designing and manipulating the systems at the molecular level, the limit of detection (LoD) of MEMSs/NEMSs and FETs rarely reaches 10−17 M (~600 copies in 100 μl) in buffer or diluted biofluid28,33,34,35,38,39,40,41,42,43,44. The sensitivity is lower in complex and high-ionic-strength biofluids, meaning these systems do not meet the requirements for precise detection and diagnosis.

In this protocol, we describe a testing platform based on a molecular-electromechanical system (MolEMS) that is manufactured and actuated with molecular-scale precision, permitting detection of trace analytes in complex biofluids. By using a graphene FET (g-FET) equipped with a MolEMS, unamplified detection of thrombin, Hg2+, ATP and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) nucleic acids has been achieved with an LoD of several copies/molecules in 100-μl buffer, serum or nasopharyngeal swab samples45. The applicability to different kinds of analytes holds promise in wide-ranging applications such as rapid pathogen screening, disease diagnosis, health monitoring and food and environmental safety.

Development of the protocolFET biosensor

An FET is an elementary unit in solid-state electronics. It typically consists of a semiconductor channel with electrodes at either end referred to as the ‘drain’ and the ‘source’. A gate electrode is placed in very close proximity to the channel through a dielectric layer. By applying a potential across the dielectric layer, an electric field modulates the conductivity of the semiconductor and controls the current flowing through the channel. Because of their inherent capability of signal amplification, FETs have emerged as candidates for sensitive bio-detection, with advantages such as fast response, label-free detection, easy integration of multiple devices for high throughput, user friendliness and suitability for point-of-care testing46.

An FET biosensor is usually configured as a liquid-gated FET (Fig. 1), where desired probes are functionalized on the surface of the channel. When exposed to an aqueous environment, an electrical double layer formed at the interface between the electrolyte solution and the channel serves as the dielectric layer47. Once charged biomolecules bind to the probes, a change in surface potential is induced, thereby altering the charge dispersal of the underlying semiconductor material. This results in a change in conductance of the FET channel, transducing the biochemical signal to electrical signal in a real-time, specific and label-free manner39,41,48,49,50,51,52,53,54.

Fig. 1: Schematic illustration of a liquid-gated FET sensor.figure 1

Probes are functionalized on the semiconducting channel. Recognition of analyte by probes induces a change in conductivity of the channel. The probe could be an antibody, single-stranded DNA (ssDNA; or aptamer), enzyme, CRISPR–Cas, etc. The analytes include proteins, nucleic acids, ions, organics and pathogens.

Challenges

Given the complexity of biological samples, sensitive detection of trace analytes in biofluids requires a carefully designed sensing interface. In particular, the sensing interface of the FET must have the ability to (i) recognize the analytes with high efficiency, (ii) resist non-specific adsorption, and (iii) overcome the Debye length limitation in high-ionic-strength biofluids. In biological detection, the Debye length may limit detection because analyte recognition events beyond this length from the FET surface are shielded by ions in solution55. Although successful efforts toward developing FET sensors that fulfil these individual features have been reported4,35,56, it is difficult to simultaneously satisfy the above requirements with the existing FET sensors. To solve this problem, the key challenge is to successfully manipulate sensing processes to achieve all the aforementioned capabilities.

DNA nanostructures in bio-detection

Single-stranded DNA (ssDNA) aptamer probes have received considerable attention because of their high specificity, affinity, stability, synthetic availability and batch-to-batch uniformity57. Thus, ssDNA probes targeting a wide range of analytes have been extensively used in FET biosensors33,58,59,60. Furthermore, the intrinsic lock-and-key assembly mechanisms61,62 and the mass commercial synthesis of nucleic acid have led to rapid progress in DNA nanotechnology over recent decades63,64. DNA sequence can be designed to construct desired conformation structures, allowing the development of different sensing mechanisms for various scenarios65,66. In particular, nucleic acid sequence can be manipulated to construct numerous DNA nanostructures with a designed configuration and dimensions via a one-step self-assemblage mechanism. DNA nanostructures serving as probe carriers enable precise spatial arrangement of biomolecules with a theoretic resolution of a single nucleotide, permitting manipulation of sensing events with molecular precision67. The high programmability of 2D and 3D DNA nanostructures has leveraged many artificial biological systems such as molecular machines68, molecular reactors69 and molecular carriers70. These DNA nanostructure-constructed systems help to not only understand fundamental mechanisms of biological processes but also manufacture functional systems including biosensors with higher precision.

From a MEMS/NEMS to a MolEMS

MEMSs/NEMSs, the miniaturized functional systems integrating electrical and mechanical components at micrometer or nanometer scale, function as an interface between microelectronic or nanoelectronic components and the environment. The micrometer- and nanometer-length scales are particularly relevant to biological materials because these scales are comparable to the size of cells, diffusion lengths of molecules and electrostatic screening lengths of ion-conducting fluids. Hence, the applications of MEMS and NEMS in bio-detection have expanded over the past decades, evolving from MEMS-based methods such as quartz crystal microbalance sensors, microfluidic sensor chips and cantilevers to NEMS-based methods such as nano-resonators71,72,73,74,75,76,77,78. By integration with FETs, MEMSs/NEMSs can be used as sensitive biosensors, because FETs combine efficient transducers with signal amplifiers in which a small parameter alteration induces a pronounced change in channel current. It has been demonstrated that the miniaturized feature size of MEMSs/NEMSs significantly improves sensing performance while reducing cost, volume, weight and power consumption79,80,81. In considering this point, further decreasing the feature size is highly desired.

Inspired by highly precise structure and versatile functions offered by DNA nanostructures, we describe a recently developed electromechanical system, named ‘MolEMS’, which reduces the feature size down to the molecular scale and resolves many of the challenges of FETs to allow for rapid detection of trace analytes in biofluids45. The MolEMS is a self-assembled DNA nanostructure containing a stiff tetrahedral double-stranded DNA (dsDNA) base and a flexible ssDNA cantilever with a probe on the tip (Fig. 2a). The base is immobilized on the channel of a liquid-gated g-FET. A severe charge screening effect exists in high-ionic-strength solutions, such as serum or blood, leading to low sensitivity of direct detection of analytes in the physiological environment. Upon electromechanical actuation by applying a negative gate voltage (Vg) on the g-FET (Fig. 2b), the cantilever can be electromechanically actuated downward, modulating sensing events close to the graphene channel. Thus, the MolEMS g-FET overcomes the Debye length limitation (Fig. 2e), permitting efficient signal transduction. Meanwhile, the high-density stiff tetrahedral bases functionalized on the channel (Fig. 2c,d) serve as a built-in antifouling layer, preventing non-specific adsorption to graphene. By manipulating the sensing operation at a molecular scale not achieved by other traditional FET sensors, the detection of trace analytes can be realized in complex biofluids such as serum and could be expanded to blood, urine and saliva.

Fig. 2: The working principle of a MolEMS g-FET.figure 2

a, Sensor components of a MolEMS that resemble a cantilever system including a rigid tetrahedral double-stranded DNA (dsDNA) base and a flexible ssDNA probe. b, Configuration of a MolEMS g-FET. The MolEMS is immobilized on the channel surface of a liquid-gated g-FET, with an Ag/AgCl electrode serving as the gate electrode. c, Optical microscope image of the graphene channel. Scale bar, 100 μm. d, Atomic force microscopy image of the MolEMS immobilized on the graphene surface. The image was obtained in 1× Tris-magnesium sulfate (1× TM) buffer. The color bar indicates the height. Scale bar, 100 nm. The measured size is consistent with the theoretical dimensions of the MolEMS. e, The working principle of the MolEMS. At a negative gate voltage (Vg), analytes recognized by the ssDNA probes (i) are actuated to the graphene surface within the Debye length (ii), while the rigid tetrahedral dsDNA bases functionalized on the graphene surface function as an antifouling layer to resist non-specific adsorption of background molecules in biofluids (iii). Panels c and d adapted from ref. 45, Springer Nature Ltd.

Comparison with other methods

Because they are affected by orders-of-magnitude higher amounts of background biomolecules or inefficient signal transduction, mainstream clinical assay approaches such as ELISA and other immunoassay approaches might be incapable of achieving an LoD of ~10−13 M82,83,84,85,86,87,88, which is required for trace analyte detection in some biological research and diagnostic applications. To overcome this challenge, analyte enrichment or amplification strategies have been used. For example, gold or magnetic nanoparticles are used to pre-concentrate analytes so that the signal can be enhanced89,90,91. Nucleic acids need be amplified by PCR17,18, loop-mediated isothermal amplification92,93,94 or recombinase polymerase amplification95,96 before measurement. However, these strategies inevitably complicate the sensing process and are time consuming. Although rapid testing methodologies based on colorimetric assays97,98, electrochemical assays4,25,99, SPR23,24 and CRISPR100,101,102 have been extensively investigated, unfortunately, the LoD rarely reaches 10−17 M in bulk buffer or diluted biofluids28,33,34,

留言 (0)

沒有登入
gif