Triple-Probe DNA Framework-Based Transistor for SARS-CoV-2 10-in-1 Pooled Testing.

Accurate and population-scale screening technology is crucial in the control and prevention of COVID-19, such as pooled testing with high overall testing efficiency. Nevertheless, pooled testing faces challenges in sensitivity and specificity due to diluted targets and increased contaminations. Here, we develop a graphene field-effect transistor sensor modified with triple-probe tetrahedral DNA framework (TDF) dimers for 10-in-1 pooled testing of SARS-CoV-2 RNA. The synergy effect of triple probes as well as the special nanostructure achieve a higher binding affinity, faster response, and better specificity. The detectable concentration reaches 0.025-0.05 copy μL-1 in unamplified samples, lower than that of the reverse transcript-polymerase chain reaction. Without a requirement of nucleic-acid amplification, the sensors identify all of the 14 positive cases in 30 nasopharyngeal swabs within an average diagnosis time of 74 s. Unamplified 10-in-1 pooled testing enabled by the triple-probe TDF dimer sensor has great potential in the screening of COVID-19 and other epidemic diseases.

Introduction

As of January 2022, the global coronavirus disease 2019 (COVID-19) pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has induced 330 million infectious cases and 550 thousand deaths. The day-by-day increasing number of infected patients puts a momentous burden on public healthcare systems, resulting in a lack of medical resources and a crash of numerous economies.[1−3] The interpersonal transmission and community spread model demands the development of easily operated diagnosis technology which can quickly identify infected individuals from a wide population for preventing the outbreak. Pooled testing with high detection throughput is the most frequently used method for population-scale screening, especially in regions with a population prevalence below 10%.[4−6] However, during the sample pooling process, the viral load of the positive sample decreases relative to the increasing sample volume when mixed with negative samples, frequently causing false-negative results.[6] The additional introduction of biomolecules and contaminations also raise the difficulty to accurately capture the targets. Thus, sensitivity and accuracy are the critical requirements in pooled testing. The quantitative reverse transcript polymerase chain reaction (qRT-PCR), as the gold-standard nucleic-acid detection technology, is the most convincing and widely employed method for COVID-19 testing. Nevertheless, it requires extraction and amplification procedures (>2 h) as well as skilled technicians and specific laboratories and equipment.[7−9] Meanwhile, other established nucleic-acid testing methodologies face challenges in sensitivity and accuracy, which increase the difficulties when testing pooled samples.[10−22] Therefore, it is urgently required to develop a rapid, sensitive, and easily operated method for SARS-CoV-2 nucleic-acid pooled testing.

The aptamer field-effect transistor (apt-FET) sensor is a pervasive method for tracing biological analytes owing to its relatively high sensitivity, specificity, and easily fabricated features.[23−26] Considerable efforts have been made to improve the sensitivity, which mainly focus on the sensing materials, the design of the DNA probes, and the exploitation of new sensing mechanisms. Hwang et al. put forward a “deformed graphene channel” strategy and significantly improved the sensitivity, which led to a limit of detection (LoD) of single-stranded DNA (ssDNA) down to 10–19 mol L–1.[27] A graphene field effect transistor (g-FET) sensor modified with multiprobe DNA nanomaterials may achieve the sensitivity requirement in pooled testing. Graphene is a single atomic-layer sensing material with high carrier mobility. The charge carriers (holes or electrons) in graphene are confined and exposed in a narrow range that is closed to the surface which makes graphene ultrasusceptive to surrounding electrostatic variation and generates low intrinsic electrical noise.[28] These combined features make graphene a sensitive material for biosensing applications.[29] However, until now, the limit of detection (LoD) has ranged from 10–15 to 10–9 mol L–1, which does not meet the sensitivity demand of pooled testing.

Creatures with multitentacle anatomies formed through thousands of years of evolution because this structure remarkably improves their sensitivity to smell and ability to catch and hunt.[30−32] Inspired by these creatures, researchers have developed many multiprobe sensors that can notably increase the capture efficiency, resulting in higher detection sensitivity and selectivity.[33−39] Nevertheless, these multiprobe sensors still cannot realize single- or few-molecule detections. Recent advances in structural DNA nanotechnology offer an accurate and controllable method to synthesize various DNA nanostructures with specific functions, which can improve the binding affinity to targets by designing different configurations in biodetection.[40−42] Here, we develop a g-FET sensor modified with a triple-probe tetrahedral DNA framework (TDF) dimer for SARS-CoV-2 RNA testing and modulate the sensing process at the molecular scale. The synergy effect of the triple probe improves the binding affinity and enables g-FET sensors to directly detect 0.025–0.05 copy μL–1 (0.4–0.8 × 10–19 mol L–1) SARS-CoV-2 RNA in artificial saliva without amplification. The sensor exhibits ∼100% overall percent agreement (OPA) with RT-PCR results within an average diagnosis time of 74 s when testing nasopharyngeal swab samples. Because of its high sensitivity and accuracy, the sensor enables 10-in-1 nucleic-acid pooled testing, showing great values in population-scale screening of COVID-19 and other infectious diseases.

Results and Discussion

The workflow of the triple-probe TDF dimer g-FET sensor for SARS-CoV-2 RNA testing is illustrated in Figure a. The sensor is configured from a liquid-gated g-FET with a polydimethylsiloxane (PDMS) chamber on the graphene channel (Figure b) with triple-probe TDF dimers on the sensing interface (enlarged diagram). Each worked sensing area (Figure c) is 30 × 150 μm2 (W × L). The designed DNA nanostructure is a probe-tunable TDF dimer with an edge length of 7 bp (∼2.6 nm of theoretical length), and then, one, two, or three probes are anchored (Figure S2a) on the top tetrahedron. All of the DNA sequences (5′–3′) are listed in Table S1 (Supporting Information). The synthesized TDF monomer and dimer nanostructures are analyzed by 10% polyacrylamide gel electrophoresis; most of the lanes only have a major band, and the yield is calculated as ∼60% (Figure S2b), which can confirm the successful synthesis of the TDF dimer nanostructure.

Figure 1

Triple-probe TDF dimer g-FET sensor for SARS-CoV-2 RNA testing. (a) Workflow and schematic diagram of the triple-probe TDF dimer g-FET sensor for SARS-CoV-2 RNA testing. The enlarged diagram is the g-FET sensing surface modified with a triple-probe TDF dimer. (b) Digital photograph of a working device. (c) Optical microscope image of the g-FET channel. (d) AFM image of the sensing surface after triple-probe TDF dimer immobilization (measured in fluid).

The TDF dimers are immobilized onto the g-FET device (Figure S3) through molecule linkers: 1-pyrenebutanoic acid succinimidyl ester (PASE).[43,44] The Raman spectrum (Figure S1a–c) demonstrates that PASE is modified on the graphene surface through π–π interaction[45,46] and remains homogeneous after PASE modification. An atomic force microscopy (AFM) image measured in fluid reflects that TDF dimers are anchored on the sensing surface in an orderly manner (Figure d). AFM images (Figure S4) measured in air show the morphology of the sensing surface, and the roughness of the surface observably increases when graphene is modified with PASE (2.0–2.5 nm) and is immobilized with triple-probe TDF dimer (5.8–7.0 nm) compared to that (0.8–1.2 nm) of the intrinsic graphene surface. Moreover, the appearances of the N 1s peak and P 2p peak (Figure S5a,b) in X-ray photoelectron spectroscopy (XPS) measurements after modification also verify that the PASE and TDF dimer are anchored on the graphene sensing surface.[46] In the transfer curve measurement of the device, the π–π interaction between PASE and graphene induces an n-doping effect on graphene, giving rise to a negative offset (8–20 mV) of VDirac (as the liquid-gate voltage Vlg when the drain-source current Ids reaches its minimum). After TDF dimer immobilization, VDirac further shifts negatively by 22–32 mV. All of these results demonstrate the successful modification of PASE and TDF dimer on the g-FET sensing surface.

The three probes designed in this work target SARS-CoV-2 RNA at the ORF1ab gene (nt 13377–13404), RdRp gene (nt 15469–15494), and E gene (nt 26332–26358) regions, respectively (Figure a). The transfer curve measurement (Ids–Vlg) of the triple-probe TDF dimer g-FET device shows the response when adding the IVT RNA solutions (Figure b); the VDirac exhibits a continues negative shift signal to ∼60 mV when the RNA solution concentration changes from 0.025 to 1000 copy μL–1. Compared to the Ids–Vlg responses of single- and dual-probe TDF dimer devices to IVT RNA (Figure S7a,b), the VDirac of the triple-probe TDF dimer presents an at least 2-fold signal enhancement. In addition, the VDirac of the triple-probe TDF dimer device exhibits a ∼10 mV negative offset when 0.025 copy μL–1 target RNA solution is added (Figure b), which demonstrates that the g-FET sensor modified with the triple-probe TDF dimer is highly sensitive.

Figure 2

SARS-CoV-2 RNA testing. (a) Triple-probe TDF dimer structure, viral Genome map, and targeted regions of three probes. (b) Transfer curve measurement of adding different concentrations of target RNA (Ids–Vg response curve). (c) Real-time |ΔIds/Ids0| response upon different concentrations of target RNA (red line, modified with triple-probe TDF dimer; gray line, without immobilized probes). (d) |ΔIds/Ids0| responses of single- and triple-probe TDF dimer g-FET sensors to different concentrations of target RNA. (e) |ΔIds/Ids0| responses of the triple-probe TDF dimer g-FET sensor to SARS-CoV-2 target RNA (0.05 copy μL–1) and nontarget RNA (samples II, III, and IV: 0.5 copy μL–1). All RNA samples were dissolved in full saliva with the addition of 12.5 mM Mg2+, and the pH is 7.0–7.4. The error bars of parts d and e are defined by the standard deviation of the results from at least 3 parallel experiments.

The real-time responses (Ids–t) of the triple-probe TDF dimer g-FET device to target RNA solution are recorded (Figure c) with different concentrations (from 0.025 to 1000 copy μL–1). Remarkable electrical response signals (|ΔIds/Ids0| decrease) are detected within 5 min even when the target RNA solution concentration is 0.025 copy μL–1, and the response time is shorter than that (10–15 min) of dual- and single-probe TDF dimer devices (Figure S7c,d). Here, we define the response values |ΔIds/Ids0| = |(Ids – Ids0)/Ids0|, where Ids is the real-time drain-source current, and Ids0 is the initial current. As for the g-FET device without immobilized probes, no obvious response signals (Figure c) are observed even when 1000 copy μL–1 target RNA solution is added (gray line). In addition, from the |ΔIds/Ids0| response values (Figure d) of the triple-probe TDF dimer g-FET sensor, the calculated LoD reaches 0.01 copy μL–1 (Figure S9). Then, the |ΔIds/Ids0| responses of the single- and dual-probe TDF dimer device to target RNA solution are also measured under the same conditions; the response values of the triple-probe TDF dimer device are at least twice that of the single-probe device (Figure d, Figure S7e). This is mainly attributed to the high probability and matching rate of the triple probe, which enables more efficient binding to target RNA and then showcases a larger |ΔIds/Ids0| response value and shorter response time in Ids–t measurements as well as a larger VDirac offset in Ids–Vlg measurements. Moreover, we measured the Ids–Vlg and Ids–t response of the device modified with ssDNA probes (Figure S8). The ssDNA probes contain three corresponding types of probes mixed at a 1:1:1 ratio. It is found that VDirac negative offset is negligible, and the |ΔIds/Ids0| response values are smaller than those of triple-, dual-, and single-probe TDF dimer g-FET devices.

Then, we measured the Ids–t response values by adding three nontargeted RNA samples with different concentrations (0.5, 5, 50, and 100 copy μL–1) in artificial saliva, including MERS-CoV RNA (sample II), SARS-CoV RNA (sample III), and human RNA (sample IV). The |ΔIds/Ids0| response values of nontargeted samples (Figure e, Figures S10 and S11) are negligible compared to that of the target sample (sample I: SARS-CoV-2 RNA solution). Therefore, we conclude that the triple-probe TDF dimer modified g-FET sensor exhibits an excellent detection performance with ultrasensitivity, short response time, and high specificity. By replacing the probes on the top tetrahedra of the triple-probe TDF dimers, we also measured the transfer curve and real-time response of the sensor upon different concentrations of cDNA (extracted from a confirmed COVID-19 case reverse transcribed cDNA). From the |ΔVDirac| response and |ΔIds/Ids0| response values (Figure S12), we find that the sensors also have excellent detection performances when testing the cDNA samples, which shows a much broader applicability of the sensor.

Biodetection includes biorecognition and signal transduction processes.[23−27,46,47] The efficient recognition rate to target RNA and the signal transduction process is the fundamental reason that the triple-probe TDF dimer g-FET sensor achieves such an excellent detection performance. The recognition upon target RNA is essentially a DNA hybridization reaction, and the binding affinity between the DNA probe and target RNA is the crucial factor in this process. The signal transduction process can be amplified via the FET device and reflected in the electrical measurement, which is mainly revealed in the VDirac offset of transfer curve test and the current change of real-time measurement. The DNA hybridization reaction can induce charge accumulation on the graphene surface and then cause a doping effect on graphene, and the monolayer graphene with high mobility enables an efficient and sensitive signal transduction process.

We chose another two types of DNA probes (Figure a,b) as control experiments and tested the Ids–Vlg responses to target IVT RNA. One probe type is the single-probe TDF dimer, and the other one is three types of ssDNA probes mixed with a 1:1:1 ratio. Meanwhile, we calculate the binding affinity by using the normalized response of ΔVDirac/ΔVDirac,max; the correlation between ΔVDirac/ΔVDirac,max and target RNA concentration is described by the Hill–Langmuir model:[48]where ΔVDirac,max is the maximum ΔVDirac, denoted ΔVDirac,max = VDirac,max – VDirac,0 (VDirac,0 and VDirac,max refer to the offset of adding the zero and maximum concentration of target RNA solution, respectively); A is the saturation response coefficient of the sensing system; and n is the Hill coefficient corresponding to the binding cooperativity. The pseudo-KD of the triple-probe TDF dimer sensor is 0.1–0.3 × 10–19 mol μL–1 (∼0.01 copy μL–1) and is calculated from the fitted curve (Figure d), which is 3 and 7 orders of magnitude lower than those of the single-probe TDF dimer sensor and ssDNA sensor, respectively. The fitted Ids–Vlg responses (Figure e) of triple- and single-probe TDF dimer and ssDNA probe modified sensors are obtained by using the standard deviation of 8 parallel measurements, and the results reveal that the triple-probe TDF dimer sensor has an offset of ΔVDirac larger than those of the other two types of sensors. The testing repeatability of the three different types of DNA probe sensors (Figure f) also indicates that the triple-probe TDF dimer sensor has a remarkable signal compared to the other two types of sensors. Furthermore, we also measured the transfer curve response and real-time test (Figures S6 and S7) of the other three control types of DNA probe sensors (dual- and single-probe TDF dimer and ssDNA probes). These results reveal that the triple-probe TDF dimer g-FET sensor exhibits a better sensing ability than the others. The synergy effect of the triple probe enables a higher binding affinity and shorter biorecognition time compared to the dual and single probe. Besides, unlike ssDNA probes, the existence of the TDF dimer structure can effectively prevent the probe DNA intertwining with each other and avoid nonspecific biomolecule adsorption on the graphene surface. Thus, the special structure of the TDF dimer as well as the triple probes, combined with efficient signal transduction of g-FET, gives rise to ultrasensitive detection.

Figure 3

Design of DNA probes. Schematic illustration of the g-FET sensing surface modified with (a) a triple-probe TDF dimer and (b) ssDNA in a 1:1:1 ratio. (c) Fluorescence image of the triple-probe Cy3-conjugated TDF dimer and ssDNA probe Cy3-conjugated on graphene. (d) ΔVDirac/ΔVDirac,max curve of the triple-probe TDF dimer sensor fitted by eq . (e) |ΔVDirac| comparison of ssDNA probes, single-probe TDF dimer, and triple-probe TDF dimer sensors. (f) |ΔVDirac| box plot comparison (device testing repeatability) of three types of probes. The error bars in parts d–f are defined by the standard deviation from at least 8 parallel experiment results.

The confocal fluorescence microscopy measurement indicates that the device modified with the triple-probe Cy3-conjugation (Cy3: a fluorescent dye Cyanine3, rather than a probe) TDF dimer has a stronger fluorescence intensity than the device modified with ssDNA Cy3-conjugation (Figure c). The main reason is that ssDNAs on the sensing surface are easily entangled with each other and tend to adsorb laterally on graphene when immersed in solution, inducing a quenching effect and then leading to a weak fluorescence intensity. However, the rigid, three-dimensional TDF can avoid the adsorption phenomenon, resulting in a stronger fluorescence. This result also demonstrates that the modification manner of the DNA probes on the graphene surface was consistent with the schematic diagram depicted in Figure a,b.

Furthermore, we tested 14 nasopharyngeal swabs (P1–P14) collected from SARS-CoV-2 positive patients with the cycle-threshold (Ct) of RT-PCR ranging from 24.9 to 40.5, 6 samples (F1–F6) from fever clinic patients, and 10 samples (H1–H10) from healthy volunteers. The real-time |ΔIds/Ids0| response upon P1 and F1 (Figure a) and a histogram of statistical responses (Figure b) from all clinical samples are also recorded. Although the nasopharyngeal swabs include large amounts of nonspecific biomolecules and contaminations, SARS-CoV-2 RNA from positive samples (P1–P14) can be detected by this sensor and showed considerable signals (|ΔIds/Ids0| > 0.8%), whereas weak signals (|ΔIds/Ids0| < 0.1%) are monitored (Table S2) when testing the negative samples (F1–F6) and healthy samples (H1–H10). The testing results of 30 nasopharyngeal swabs exhibit 100% OPA with RT-PCR results, as well as 0.93 sensitivity and 1.00 specificity (Figure c), indicating that this triple-probe TDF dimer g-FET sensor can clearly identify the positive and negative samples. Moreover, the response time for diagnosing COVID-19 positive samples (P1–P14) is shortened to 1–4 min with an average of 74 s (inset of Figure a) with the |ΔIds/Ids0| response value reaching 3 times that of negative patients (Figure S13a). The |ΔIds/Ids0| responses (Figures S13b,c and S14) upon other clinical samples (P3–P4, F3–F6, and H1–H10) are also measured and recorded in the Supporting Information. In addition, to verify that this sensor is more sensitive than the qRT-PCR assay, we tested the Ids–t response of the sensor to diluted clinical samples (P8: Ct = 30, diluted from 1/10 to 1/107 times) sequentially (Figure d, Figure S15). The result reveals that this sensor still achieves an observable |ΔIds/Ids0| response (0.54%) even when the sample is diluted by 1/107 times, which indicates that the triple-probe TDF dimer g-FET sensor has higher sensitivity than the qRT-PCR assay. Compared with the commercial COVID-19 detection kits[8] (Table S4) and other established methodologies, such as qRT-PCR,[9,49,50,52] the US CDC and China National Medical Products Administration (NMPA)-approved qRT-PCR[3,51] assay, the reverse transcription loop-mediated isothermal amplification (RT-LAMP)[12,53−55] assay, the clustered regularly interspaced short palindromic repeats (CRISPR)[13,14,56] assay, the electrochemical (EC)[15−17] assay, and other methods,[10,11,46,57−59] this triple-probe TDF dimer g-FET sensor assay exhibits a shorter response time and higher LoD in COVID-19 viral RNA detection (Figure e, Table ).

Figure 4

Clinical validation of the triple-probe TDF dimer g-FET sensor. (a) Real-time |ΔIds/Ids0| response of F1 and P1. The inset is the diagnosis time for P1–P14. (b) |ΔIds/Ids0| response upon an addition of clinical samples (P1–P14, F1–F6, and H1–H10). (c) Confusion matrix summarizing the assay discrimination performance between positive and negative swab samples. (d) Real-time |ΔIds/Ids0| response upon various diluted concentrations (1%, 10%, and 100%) of P8. (e) Comparison of this assay with other reported methods for SARS-CoV-2 RNA detection. (f) Workflow illustration for the 10-in-1 pooled sample testing strategy. (g) Real-time |ΔIds/Ids0| response of the 10-in-1 pooled negative sample (M1) and positive sample (M9). (h) |ΔIds/Ids0| response upon the addition of clinical 10-in-1 pooled samples (negative, M1–M7; positive, M8–M14).

Table 1

Existing COVID-19 Detection Methods, TDF Dimer g-FET, and Their Performance

detection method	analyte type	target	sample or medium	amplification	LoD	response time	clinical validation	pooled testing	ref
qRT-PCRa	ORF1ab and N gene cloned into plasmids	ORF1b and N gene	sputum samples	yes	10 copy/reaction	75 min	yes	no	(49)
qRT-PCR	viral RNA	RdRp1, RdRp2, E gene, N gene, N1, N2, and N3	nasopharyngeal swabs	yes	5 copy/μL N3; 10 for N1, N2, E; 50 for RdRp1, RdRp2	60–90 min	yes	no	(50)
qRT-PCR (China NMPA)	viral RNA	ORF1ab and N gene	nasopharyngeal swabs	yes	0.6–3.2 copy/μL	>120 min	yes	no	(3)
qRT-PCR (US CDC)	viral RNA	N1, N2, and N3	nasopharyngeal swabs	yes	1–3.2 copy/μL	>120 min	yes	no	(51)
384 RT-PCR method	viral RNA	N1, N2 gene	nasopharyngeal swabs	yes	5 copy/μL	73.2 min	yes	no	(52)
qRT-PCR	viral RNA	RdRp/helicase (Hel), S, and N gene	nasopharyngeal swabs	yes	11.2–21.3 copy/reaction	∼60 min	yes	no	(9)
iLACO (RT-LAMP)b	synthetic RNA	ORF1ab gene	nasopharyngeal swabs	yes	10 copy/reaction	15–40 min	yes	no	(53)
early detection RT-LAMPc	viral RNA	N gene	nasopharyngeal swabs	yes	118.6 copy/reaction	30–40 min	yes	no	(54)
RT-LAMP/Cas12 DETECTR assay	synthetic RNA	E and N gene	nasopharyngeal swabs	yes	10 copy/μL	45 min	yes	no	(12)
AIOD-CRISPR-Cas12ad assay	viral RNA	N gene	nasopharyngeal swabs	yes	5 copy/μL	20–40 min for extraction; 20 min for reaction	yes	no	(14)
RT-LAMP	viral RNA	ORF1ab, E and N gene	throat swab specimens	yes	1 copy/μL	30 min	yes	no	(55)
RPA/SHERLOCKe assay	synthetic RNA	S and ORF1ab gene	Hybri-detect assay buffer	yes	10 copy/μL	60 min	no	no	(56)
CRISPR/SHERLOCK assay	viral RNA	S, N and ORF1ab gene	nasopharyngeal and throat swabs	yes	42 copy/μL	>60 min	yes	no	(13)
RT-RAAf assay	viral RNA	S, ORF1ab gene	nasopharyngeal swabs	yes	10 copy/reaction for S gene	20–25 min	yes	no	(10)
Exo-IQ-RT-RPAg assay	viral RNA	N gene	nasopharyngeal swabs	yes	7.74 copy/μL	20–25 min	yes	no	(11)
electrochemical DPVh	viral RNA	ORF1ab, N gene	saliva or nasopharyngeal swabs	no	0.2 copy/μL	190 min	yes	no	(17)
electrochemical DPV	viral RNA	S, N gene	nasopharyngeal swabs	yes	1 copy/μL	<120 min	yes	no	(15)
current–voltage electrochemical assay	viral RNA	N gene	nasopharyngeal swabs	no	6.9 copy/μL	30 min for extraction, 5 min for detection	yes	no	(16)
ELISAi	antibody	human IgM and IgG	serum	no	N.A.j	60–180 min	yes	no	(57)
chemiluminescent immunoassay	antibody	human IgM and IgG	reaction mixture	no	4.6 μM	48 min	no	no	(58)
MALDI-MSk	virus antigen	SARS-CoV-2 S-protein	nasopharyngeal swabs	no	N.A.	>20 min	yes	no	(18)
FET-based biosensor	antigen	SARS-CoV-2 S-protein	nasopharyngeal swabs	no	0.242 copy/μL	>1 min	yes	no	(46)
electrochemical DPV	antigen	SARS-CoV-2 S-protein	saliva	no	16.666 copy/μL	10–30 s	no	no	(59)
TDF dimer g-FET assay	viral RNA	ORF1ab, RdRp and E gene	nasopharyngeal swabs	no	0.025–0.05 copy/μL	1–4 min (74 s in average)	yes	yes	this work

qRT-PCR: Quantitative reverse transcription-polymerase chain reaction.

iLACO (RT-LAMP): Isothermal LAMP-based method for COVID-19.

RT-LAMP: Reverse transcription loop-mediated isothermal amplification.

AIOD-CRISPR-Cas12a: All-in-one dual clustered regularly interspaced short palindromic repeats Cas12a.

RPA/SHERLOCK: Recombinase polymerase amplification/specific high sensitivity enzymatic reporter UnLOCKing.

RT-RAA: Reverse transcription recombinase-aided amplification.

Exo-IQ-RT-RPA: Exoprobe with an internally linked quencher reverse transcription recombinase polymerase amplification.

DPV: Differential pulse voltammetry.

ELISA: Enzyme linked immunosorbent assay.

N.A.: Not available.

MALDI-MS: Matrix-assisted laser desorption/ionization mass spectrometry.

The above testing results demonstrate that the sensor exhibits a rapid and sensitive detection capability, which can solve the extrinsic problems of having insufficient sensitivity and being a time-consuming process, thus enabling 10-in-1 pooled testing. The workflow of pooled testing is illustrated in Figure f. First, 200 μL portions of each nasopharyngeal swab sample are collected from 10 healthy individuals and then mixed in one pool as the negative samples (M1–M7). For the 10-in-1 positive pooled samples (M8–M14), 200 μL portions of each nasopharyngeal swab sample are collected from 9 healthy individuals and mixed with 200 μL portions of each nasopharyngeal swab sample from positive samples (P8–P14) in one pool, respectively. The real-time |ΔIds/Ids0| responses of M1 and M9 (Figure g) and M2–M4, M8, and M10 (Figure S16) are also recorded by this sensor. All of the samples (M8–M14) mixed with one positive sample exhibit larger |ΔIds/Ids0| responses from 2.69% to 10.58%, while the negative mixture samples (M1–M7) only generate weak responses from 0.056% to 0.228% (Figure h, Table S3). These 10-in-1 pooled testing results reveal that the sensor can rapidly and remarkably distinguish the pooled samples mixed with 1 positive nasopharyngeal swab. This sensor provides an accurate and easily operated detection method that exhibits great practical value in rapid identification of COVID-19 infected individuals from a wide population and in realizing large-scale screening of infectious epidemic diseases.

Conclusions

In this work, we develop a g-FET sensor modified with a triple-probe TDF dimer that realizes unamplified testing of samples from COVID-19 infected individuals as well as 10-in-1 pooled samples. Compared with existing technologies, this sensor achieves shorter diagnosis time and higher sensitivity via improving the binding efficiency by the synergy effect; it overcomes extrinsic defects when employed in 10-in-1 pooled testing. The sensor can identify the positive pooled samples despite its low viral load, which reveals great value in solving the problem of point-of-care detection and population-scale screening for COVID-19, as well as the mutated strains of SARS-CoV-2 virus and other infectious diseases via replacing the probes. Moreover, the sensor can be developed into a comprehensive platform when integrated with a portable microelectronic system, which can rapidly and accurately monitor the COVID-19 patients in airports, rail stations, and other public places with huge visitor flow rates. Rapid, portable, easily operated 10-in-1 nucleic-acid pooled testing, which increases the testing throughput and reduces the consumption of medical resources, can alleviate the burden on public healthcare services.