Highly-specific aptamer targeting SARS-CoV-2 S1 protein screened on an automatic integrated microfluidic system for COVID-19 diagnosis

Abstract

Variants of the severe acute respiratory syndrome coronavirus (SARS-CoV-2) have evolved such that it may be challenging for diagnosis and clinical treatment of the pandemic coronavirus disease-19 (COVID-19). Compared with developed SARS-CoV-2 diagnostic tools recently, aptamers may exhibit some advantages, including high specificity/affinity, longer shelf life (vs. antibodies), and could be easily prepared. Herein an integrated microfluidic system was developed to automatically carry out one novel screening process based on the systematic evolution of ligands by exponential enrichment (SELEX) for screening aptamers specific with SARS-CoV-2. The new screening process started with five rounds of positive selection (with the S1 protein of SARS-CoV-2). In addition, including non-target viruses (influenza A and B), human respiratory tract-related cancer cells (adenocarcinoma human alveolar basal epithelial cells and dysplastic oral keratinocytes), and upper respiratory tract-related infectious bacteria (including methicillin-resistant Staphylococcus aureus, Pseudomonas aeruginosa, Acinetobacter baumannii, and Klebsiella pneumoniae), and human saliva were involved to increase the specificity of the screened aptamer during the negative selection. Totally, all 10 rounds could be completed within 20 hr. The dissociation constant of the selected aptamer was determined to be 63.0 nM with S1 protein. Limits of detection for Wuhan and Omicron clinical strains were found to be satisfactory for clinical applications (i.e. 4.80 × 10¹ and 1.95×10² copies/mL, respectively). Moreover, the developed aptamer was verified to be capable of capturing inactivated SARS-CoV-2 viruses, eight SARS-CoV-2 pseudo-viruses, and clinical isolates of SARS-CoV-2 viruses. For high-variable emerging viruses, this developed integrated microfluidic system can be used to rapidly select highly-specific aptamers based on the novel SELEX methods to deal with infectious diseases in the future.

Graphical abstract

An integrated, automatic microfluidic device for screening of aptamers against SARS-CoV-2 S1 proteins was developed. The screened aptamer with a high affinity and specificity was capable to SARS-CoV-2 viruses and could be used for diagnosis of COVID-19.

1. Introduction

The severe acute respiratory syndrome coronavirus (SARS-CoV-2) is a single-stranded RNA virus, which belongs to the betacoronavirus subfamily of the coronaviridae family; the viral genome is around 30 kb [1,2]. SARS-CoV-2 has several major open reading frames (ORFs), including ORF1a, ORF1b, and some structural proteins. Among them, ORF1a/b is responsible for whole genome synthesis and replication and is translated into 16 non-structural proteins (e.g., RNA-dependent RNA polymerase). Moreover, envelope (E), nucleocapsid, spike proteins (S), and membrane proteins are critical for assembling intact viral particles [[3], [4], [5]]. The virus binds human angiotensin-converting enzyme 2 (ACE2) with a receptor binding domain (RBD) of the S protein, which allows it to fuse with the cell membranes, enter the cells, and proliferate within cytosol [6,7]. The S protein, which is composed of S1 and S2 subunits, is considered a reliable diagnostic biomarker of COVID-19 [[8], [9], [10]].

Coronavirus disease-19 (COVID-19), which is caused by SARS-CoV-2, is regarded as the greatest public health concern since 2019; despite the widespread availability of vaccines, many are still succumbing to this illness. The World Health Organization (WHO) reported 660 million cases and 6.69 million deaths, respectively, as of January 2023. The high mutation rates of the viruses, particularly in their spike proteins, likely mean that vaccines may be under constant development for years to come [[11], [12], [13]]. These SARS-CoV-2 variants are monitored by the Centers for Disease Control and Prevention, and are classified by the SARS-CoV-2 Interagency Group in the United States as “variants being monitored,” of which they are then deemed either “variant(s) of interest” or “variants of concern” [14,15]. Till now, the number of variants continues to rise. Accordingly, the high mutation of the spike protein may enhance SARS-CoV-2 to easily escape the attack of the immune system and vaccines [16].

For the outbreak of these emerging viruses, there are two common methods for detection currently. Quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) is the gold standard for detecting specific genes of SARS-CoV-2 due to its high sensitivity and specificity [17]. However, it requires 3–4 hrs for sample preparation (including virus lysis and RNA extraction) and nucleic acid amplification for the entire diagnostic processes. In addition, viral antigen or immunoglobulin G and M (IgG/IgM) antibody serological kits are also used to simply and rapidly detect SARS-CoV-2. However, sensitivity and specificity are relatively low when compared with nucleic acid techniques [18,19].

Recently, single-stranded DNA (ssDNA) aptamers have been developed against some human pathogens, including SARS-CoV-2 [[20], [21], [22], [23]]. These nucleic acids can fold into specific 3D structures by sequences that allow them to function as effective probes with targets. The specific aptamer was screened from random libraries of nucleic acids by systematic evolution of ligands by exponential enrichment (SELEX) [24,25]. For instance, aptamers screened against SARS-CoV-2 target the RBD region of the S1 subunit have been reported recently [20,26,27]. However, the conventional SELEX requires a lengthy screening process (up to months), which may not fast enough for virus mutation; by the time a suitable aptamer probe is identified, the virus may already have mutated. For instance, since the initiation of the Wuhan strain in 2019, variants from Alpha, Beta, Delta to Omicron (such as BA.4, BA.5, and BF.7) were successively isolated in the past two to three years [28,29]. Therefore, rapid chip-based SELEX methods, such as capillary electrophoresis (CE)- [30] and microfluidic-based SELEX [31] approaches had developed to rapidly select favorable aptamers against highly mutant pathogens [26,31]. However, selected ssDNA still required extra on-bench polymerase chain reaction (PCR) processes for most chip-based SELEX [26,31]. Till now, an integrated microfluidic device to screen aptamers specific to SARS-CoV-2 S1 spike protein has never been reported in literature. In this study, a new, fully-automated microfluidic system capable of rapidly screening aptamer probes for the S1 protein of SARS-CoV-2 within a novel ten rounds of selection, including five positive selections, two competitive selections, and three negative selections to select high-affinity, high-specificity aptamers targeting S1 proteins of SARS-CoV-2. Pre-programmed steps such as reagent transport, on-chip PCR, and magnetic beads collection using a new magnetic control module was developed for automating screening of high-affinity, highly-specific aptamers against SARS-CoV-2 viruses within 20 h.

More importantly, since viral RNA was reported to exist in respiratory clinical samples for more than 1 month after onset even though the viruses cannot be cultured after the third week of infections [32], some patients may be rep0rted as false-positive results after qRT-PCR since it could still amplify RNA residues. Therefore, avoiding or decreasing these false-positive results plays an important role in COVID-19 diagnosis. With a highly-specific aptamer specific to SARS-CoV-2 S1 to capture virus particles (not viral RNA), false-positive cases could be alleviated.

2. Materials and methods

2.1. Preparation of reagents, ssDNA library and SARS-CoV-2 S1-coated magnetic beads

Ten μL of ssDNA library (10 μM) from Medclub Scientific (Taiwan) represented a pool of randomly generated 72-bp oligonucleotides. Each ssDNA contained two 16-bp primer-annealing regions, with 40 random nucleotides in the middle: 5′-GGCAGGAAGACAAACA–random 40 bp-TGGTCTGTGGTGCTGT-3'. A recombinant, 681 amino acids, 76.5 kDa SARS-CoV-2 S1 subunit (His-tagged, Sino Biological, China) was coated onto HisPur™ nickel-nitriloacetic acid (Ni-NTA) magnetic beads (12.5 mg/mL, ɸ = 1 μm, Thermo Fisher Scientific [TFS], USA) as follows. First, 10 μL of the beads were washed twice with 200 μL of equilibration buffer (100 mM sodium phosphate, 600 mM sodium chloride, 0.05% Tween™-20 Surfact-Amps™, and 30 mM imidazole). After washing, these beads were resuspended in 80 μL of equilibration buffer. Then, 20 μL of S1 proteins (250 mg/L) were added and incubated for 30 min at 20 RPM under the C1 setting of an ELMI Intelli-mixer (ELMI Ltd., Latvia). Next, the mixture was placed on a DynaMag™-2 magnetic rack (TFS) for 2 min to collect the S1-coated beads, which were then washed twice with 200 μL of equilibration buffer. Finally, the S1-coated magnetic beads were resuspended in 100 μL of equilibration buffer and stored at 4 °C until use.

2.2. On-chip SELEX

On-chip SELEX screening of the S1 protein (Fig. 1 ) was carried out by an unique 10 selection rounds, including five positive, two saliva-competitive, and three negative selection rounds, that were automatically performed (Fig. 2 a) on a microfluidic system composed of (1) a microfluidic control module for control of microvalves, micromixers, and micropumps, (2) a temperature control module for reagent preservation, denaturation, ssDNA regeneration, and PCR thermocycling, and (3) a magnetic control module for bead capture. A dual thermoelectric (TE)-cooler module featured both a “cold region” for reactant preservation and ssDNA regeneration and a “hot region” for thermocycling and denaturation. Reaction conditions were set and tracked in real-time via Bluetooth.

Fig. 1.

Schematic illustration of the SELEX screening of aptamers targeting the SARS CoV-2 S1 protein on an integrated microfluidic device. Ten rounds of selection were carried out: five for positive selection, two for competitive selection (saliva), and three for negative selection (non-target microbes & two eukaryotic cell types). Inf = influenza.

Fig. 2.

(a) The microfluidic system for automated SELEX. (b) The chip was equipped with two micromixers, two suction outlets for wastes, six micropumps, seventeen microvalves, and microchambers for the ssDNA library, binding buffer, 1x PBS, wash buffer, ssDNA regeneration, storage, protein release, and PCR reagents.

For positive selection, the ssDNA library (10 μL), PBST wash buffer (500 μL of 1×phosphate buffered saline (PBS) with 0.01% Tween-20, pH 7.4), binding buffer (40 μL), SARS-CoV-2 S1 protein-coated beads (10 μL), and PCR reagents (30 μL) were pre-loaded into the corresponding chambers of the chip (Fig. 2b). Then, 40 μL of 1× PBS (in the binding buffer chamber) was pumped into the ssDNA library chamber by activating micropump #1, and the ssDNA library was denatured at 90 °C for 10 min. Next, the denatured ssDNA was transported to the ssDNA regeneration chamber by activating the long-type micropump #6 and cooled at 8 °C for 5 min for single-strand DNA regeneration (ssDNA regeneration). Then, this treated ssDNA was transported to micromixer #2 and incubated with SARS-CoV-2 S1 protein-coated beads (10 μL) for 20 min such that few ssDNA could be captured. Note that the denatured DNA (i.e. the single strand of the aptamer) from the PCR products (i.e. double-stranded DNA) under SELEX processes should be kept at 8 °C for at least 5 min. According to a previous paper [33], it reported that the aptamer could capture the target within 10 min by a microfluidic chip. Therefore, the single-strand aptamer could be re-formed at room temperature for 20 min in the next step, which was the protocol we followed in this work. Next, the magnetic control module was used to collect the bead-ssDNA complexes. ssDNA-bound beads were washed twice with wash buffer (60 μL) for 1 min, and the unbound ssDNA was removed by suction. Then, the ssDNA-bound beads were resuspended in 20 μL of 1×PBS and mixed for 1 min. These beads were then transported into the protein release chamber located in the hot region of thermoelectric-cooling module for specific ssDNA released via thermolysis (90 °C). The magnetic control module was next used to collect the beads, and the supernatant was transported to the storage chamber. Next, 3–4 μL of collected supernatant were transported to the PCR chamber for DNA amplification as follows: initial denaturation at 90 °C for 5 min, followed by 25 cycles of denaturation at 88 °C for 30 s, annealing at 60 °C for 15 s, and extension at 72 °C for 30 s; a final elongation step at 72 °C for 5 min was then conducted. All PCR reagents included 0.5 μL of forward primer (5′-GGCAGGAAGACAAACA-3′, 10 μM), 0.5 μL of reverse primer (5′- ACAGCACCACAGACCA -3′, 10 μM, both from Genomics, Taiwan), 0.6 μL of deoxynucleoside triphosphates (10 mM, TFS), 0.2 μL of DNA polymerase (Super-Therm Gold DNA polymerase, JMR-851, Bersing Technology, Taiwan), 3 μL of reaction buffer (10× Super-Therm Gold buffer, JMR-470), and double-distilled water (ddH₂0) to 30 μL. In addition, 10 μL of mineral oil (M8410, Sigma-Aldrich, USA) were loaded to cover the reagents to avoid evaporation. Afterward, 10 μL of PCR products were reloaded into the ssDNA library chamber for the next round of SELEX. The remaining products could be used to confirm the accuracy and efficiency of the selection rounds with gel electrophoresis (described below). For each on-chip SELEX round, all samples and reagents were manually made fresh and loaded into the corresponding chambers at the beginning. All on-chip processes including positive, competitive, and negative selections for each round were fully automated.

2.3. Saliva-competitive selection and negative selection

To ensure that the screened aptamers may exhibit high affinity in a complex environment (such as saliva) featuring many off-target molecules, two rounds of saliva-competitive selection (i.e. rounds 6 and 7) were conducted. Note that the salvia sample was provided by a single donor. Aptamer affinity can vary across pH values and levels of certain ions so that these tests were especially critical [34]. Therefore, after five rounds of positive selection, the buffer was changed to human saliva (40 μL) for two rounds of saliva-competitive selection. Afterwards, common respiratory tract infection-associated viruses, bacteria, and cells were used as non-specific targets in three rounds of negative selection to further enhance the specificity of the screened aptamers. Since respiratory infections from influenza A, influenza B, and some microorganisms (such as bacteria) have similar initial oropharyngeal symptoms, the first round of negative selection (i.e. round 8) included the former two viruses. A universal aptamer for influenza A (H1N1 & H3N2) and influenza B viruses was used to immobilize these viruses on the surfaces of magnetic beads [35]. Before the first negative selection round, these carboxylic magnetic beads (4×10⁸ beads/mL, 10 μL, Φ = 1 μm, Invitrogen) which coated previously-screened aptamer were bound with influenza A H1N1 (32 hemagglutinating unit (HAU), 10 μL), influenza A H3N2 (16 HAU, 10 μL), and influenza B (16 HAU, 10 μL); after binding for 30 min and washing twice with 40 μL of PBST (1x PBS + 0.01% Tween20), they were resuspended in 40 μL of 1× PBS [35].

The round-7 SELEX products (10 μL) and bead-virus complexes (40 μL) were loaded in micromixer #1 and mixed at 1 Hz for 20 min. Then, non-target-bound beads were collected, and the supernatant was transported to the ssDNA library chamber as the round 8 target sample; note that one positive round was performed to enrich the target sample after round 8.

Round 9 was performed with four bacteria commonly found in the upper respiratory tract: methicillin-resistant Staphylococcus aureus (MRSA), Pseudomonas aeruginosa, Acinetobacter baumannii, and Klebsiella pneumoniae (all from the Department of Medical Laboratory Science & Biotechnology, Asia University, Taiwan). First, bead-bacteria complexes were prepared as follows. Protein A beads (50 μL, 20 mg/mL, diameter = 500 nm, So-Fe Biomedicine, China) were used to bind the probe targeting the flexible neck regions of mannose-binding lectin (Fc-MBL protein, 150 mg/L, Sino Biological, China) of the target bacteria; bacterial mixtures (50 μL, 10⁹ colony forming Unit, CFU/mL) were bound and washed twice with PBST prior to use, and resuspended in 40 μL of 1×PBS for negative selection [36]. Before the second negative selection round began, the selected mixture from round 8 (10 μL) and bead-bacteria complexes (40 μL) were loaded in micromixer #1, and the same on-chip process as for round 8 was enacted.

Due to regulatory limitations and the fact that the normal human cells are not easy to culture, nasal cancer and lung cell lines were used in the negative selection to exclude interference from human cells during the aptamer screening process of SELEX in this study. The third round of negative selection (i.e. round 10) included these cells that can be found in the upper respiratory tract, including adenocarcinoma human alveolar basal epithelial cells (A549, ATCC CCL-185) and dysplastic oral keratinocyte (DOK) cells (ATCC CVCL_1180; both provided by the Institute of Medicine of Chung Shan Medical University, Taiwan). Bead-cell complexes were prepared as follows. EpCAM antibody beads (10 μL, TFS) and cell line mixtures (50 μL, 10⁷ CFU/mL) were mixed, washed twice with PBST, and resuspended in 40 μL of 1×PBS. Then, 10 μL of the resulting mixture of round 9 was mixed with 40 μL of bead-cell complexes as for the previous steps.

2.4. TA cloning and aptamer sequencing

After aptamer selection, selected oligonucleotides were cloned into ampicillin-resistant TOPOTM TA vectors (Invitrogen) and transformed into E. coli as our previous study [37]; 4 μL of final amplified product were incubated with 1 μL of TOPO cloning vector and 1 μL of salt solution from the cloning kit for 5 min at room temperature [37]. Then, all mixtures were transformed into 100 μL of ECOSTM E. coli competent cells (Yeastern Biotech, Taiwan) at 42 °C for 1 min. Next, these mixtures were plated on lysogeny broth plates with 100 μM/mL of ampicillin for 16–20 hr at 37 °C. One hundred randomly selected clones were screened by PCR as for the SELEX products, and confirmed colonies were sequenced by Genomics. The 2D structure and stability of the screened aptamers were simulated from the resulting sequences on http://www.nupack.org/.

2.5. Chip design and microfabrication

For conventional SELEX selection, all processes were manually operated in each round. In this work, all fluid manipulation was achieved automatically on-chip herein. The 44 × 71 mm (length × width) chip (Fig. 2b) consisted of two polydimethylsiloxane (PDMS) layers and a glass substrate. It featured six micropumps (large (#1–4), small (#5) and long-type micropump #6), seventeen microvalves, two micromixers, eight microchambers, and two suction inlets for wastes. An open-type micromixer was used for reagent binding and mixing, and a micropump enacted precise reagent transport [36]. Detailed information on the operation of these micro-components can be found in our previous works [37,38].

The manufacturing of the chip included (1) poly-methyl methacrylate (PMMA) micromachining, (2) PDMS replication, and (3) PDMS/glass bonding [31]. First, a computer-numerical-control machine (EGX-400, Roland, Japan) was used to fabricate PMMA (Centenary Material, Taiwan) molds with microstructures for the air control and liquid channel layers. Then, after de-molding the cured PDMS layers from the molds, oxygen plasma treatment (FC-12064, Femto Science, USA) was used to bond the air control and liquid channel layers onto a glass substrate (G-Tech Optoelectronics, Taiwan) to fabricate the tri-layer pneumatic microfluidic chip [31].

2.6. Determination of the dissociation constant (K_d) of the screened aptamer

To determine affinity of the screened aptamers, various concentrations of carboxyfluorescein (FAM)-labelled aptamers (fluorescein amidite, Protech, Taiwan) were used to bind SARS-CoV-2 S1-coated beads (12.5 mg/mL). A negative control [ddH₂O] and two-fold serial dilutions of FAM-labelled aptamer (800, 400, 200, 100, 50, 25, 12.5, 6.25 and 0 nM) were used for testing. First, 50 μL of FAM-labelled aptamer were mixed with 5 μL of SARS-CoV-2 S1 protein-coated beads for 1 hr at 20 RPM (same Intelli-Mixer conditions as above). After two washes with 100 μL of 1×PBS with 0.01% Tween-20 (pH 7.4), samples were resuspended in 100 μL of 1×PBS, and the fluorescence intensities were analyzed by an FLUOstar Omega plate reader (BMG LabTech, Germany). Additional details can be found in supplemental information table 1 ( Table S1 ). The K_d value was derived by fitting the relationship between the aptamer concentration (X) and the fluorescence intensity (Y) in GraphPad Prism software (ver. 5.0, USA) as follows [39]:

Y = B_max *X / (K_d + X) + NS*X + background (1)

where B_max is the maximum specific binding and NS is the slope of the nonspecific binding in Y/X units. “Background” is the amount of nonspecific binding with no added ligand.

Moreover, the dissociation constant of the screened aptamer was also confirmed by SPR [21]. Briefly, aptamers with FAM labels were acquired commercially for SPR experiments (Genomics, New Taipei City, Taiwan). First, the aptamers were diluted to 1 μM in PBS/Mg buffer (1X PBS with 1 mM MgCl₂). The diluted aptamers were heated to 95 °C for 10 min before being immediately cooled to 4 °C for 15 min. The aptamers were then brought to room temperature for 15 min. Following that, the aptamers were diluted with PBS/Mg to the concentrations required for further studies.

The method of the SPR study was based on prior works [21,40] with minor modifications. The Biacore 3000 instrument (Cytiva US, Marlborough, MA) was used to characterize the binding affinity of chosen aptamers for this work. The CM5 sensor chips (Cytiva, US) were activated for 7 min at a flow rate of 10 μL/min with 400 mM of 3-(N, N-dimethylamino) propyl-Nethylcarbodiimide (EDC) and 100 mM of N-Hyroxysuccinimide (NHS). Recombinant S1 proteins (SARS-CoV-2 spike S1-his recombinant protein, Cat: 40591-V08B1, Sino Biological, China) were diluted to a concentration of 20 μg/mL in 10-mM sodium acetate (pH 4.5) and immobilized by amine coupling to a target level of 5000–6000 RU onto the assay channel. In the meanwhile, an empty flow cell would act as a reference channel. The remaining non-reacted functional groups on CM5 chips were blocked for 7 min at a flow rate of 10 μL/min with a 1 M ethanolamine hydrochloride (pH 8.5) solution. The sensor chip was then primed three times and equilibrated with PBS/Mg until a stable baseline was reached. For the SPR measurements, several dilutions of individual aptamers with final concentrations ranging from 0-500 nM were injected consecutively, beginning with the lowest concentration. The flow rate (30 μL/min) was kept constant throughout the entire procedure, and the association and dissociation times were retained at 2 and 5 min, respectively. The regenerate process was carried out by injected 5 M NaCl for 1 min.

2.7. Binding with SARS-CoV-2 pseudo-viruses

Pseudo-viruses were used to explore the ability of the screened aptamer to capture various kinds of coronaviruses. The pseudo-viruses were provided by the National RNAi Facility at Academia Sinica (Taiwan). Each type of SARS-CoV-2 pseudo-virus expressed specific S proteins around vector surfaces, and internal enhanced green fluorescent proteins were used to track the fluorescence signals as proxies for aptamer capture capacity. Eight pseudo-viruses were tested, including the original virus (YP_009724390), D614 (Wuhan), B.1.1.7 (Alpha), B.1.351 (Beta), Lineage P1 (Gamma), AY.1 (Delta), BA.1.1 (Omicron), and BA.4/BA.5 (Omicron). Different titters of the pseudo-viruses were diluted to 300 transducing units (TU)/μL. The conjugation method for the selected S1 aptamer onto carboxylic magnetic beads (Dynabeads® MyOne™ Carboxylic Acid, Invitrogen Co., USA) followed a protocol reported in a previous study [35]. The details were provided in the Supplemental Method M1. Briefly, twenty μL of SARS-CoV-2 S1 aptamer-conjugated beads were incubated with 30 μL of each pseudo-virus for 30 min. After washing twice with 100 μL of 1×PBS, samples were resuspended in 10 μL of 1×PBS and observed under a BX43 fluorescence microscope (Olympus, Japan).

2.8. Binding with SARS-CoV-2 inactivated virus particles

Carboxylic magnetic beads (4×10⁸ beads/mL, 100 μL, Φ = 1 μm, Invitrogen) were coated with 5′-amine-modified SARS-CoV-2 aptamer to capture inactive virus particles containing the entire genome (SARS-CoV-2 Q Control, 10⁴ copies/mL; Qnostics, UK). Ten μL of SARS-CoV-2 S1 aptamer-conjugated beads (details provided in the supplemental methods) were incubated with 100 μL of the inactivated viruses for 1 hr at 20 RPM on the Intelli-mixer (as above). After washing twice with 200 μL of 1×PBS, probe-virus-bead complexes were resuspended in 5 μL of 1×PBS and incubated at 95 °C for 10 min. Then, 5 μL of the resulting lytic viral sample was used in RT-PCR as follows: 50 °C for 5 min, 42 °C (reverse transcription) for 10 min, 95 °C (enzyme inactivation) for 3 min, followed by 45 cycles of denaturation at 95 °C for 15 s and annealing and extension at 58 °C for 1 min; the last step was the final elongation step at 58 °C for 5 min. RT-PCR reagents are listed in Supplemental Material Table S2.

2.9. Specificity tests of the screened aptamer

To further ensure the specificity of the screened aptamer, 10 μL of SARS-CoV-2 S1 aptamer-conjugated beads were incubated with 100 μL of influenza A (H1N1: 32 HAU; H3N2; 16 HAU), influenza B (16 HAU), A549 cells (10⁴ cells), DOK cells (10⁴ cells), and all four tested bacteria (10⁴ CFU) for 1 hr at 20 RPM. Next, bead-non-target complexes were collected and washed thrice with 200 μL of 1×PBS. Then, collected beads were resuspended in 5 μL of 1×PBS, lysed at 95 °C for 10 min, and PCR-amplified with the primers of Supplemental Material Table S3.

2.10. Binding with clinical isolates of SARS-CoV-2 viruses

To test the binding ability of the screened aptamer with live SARS-CoV-2 viruses, clinical isolates of SARS-CoV-2 viruses from the Department of Pathology of National Cheng Kung University Hospital (Taiwan) were used under the guidelines on biosafety level 3. We specifically tested the initial Wuhan strain and Omicron BA 5 variants (IRB number, B-ER-111-172) to verify clinical bio-applicability. Briefly, 10 μL of SARS-CoV-2 S1 aptamer-conjugated beads were incubated with 100 μL of 10-fold serially diluted viruses (1,10¹ x, 10²x, 10³x, 10⁴x, 10⁵x, 10⁶x, 10⁷x and 10⁸x-diluted) for 1 hr at 20 RPM. After two washes with 200 μL of 1×PBS, samples were resuspended in 110 μL of ddH₂O, and viral RNA was extracted with the Qiagen viral RNA mini kit (Germany). In the resulting 60 μL of viral RNA in ddH₂O, 5 μL was used in RT-PCR as described above. PCR products were analyzed on 2% tris-borate-ethylenediaminetetraacetic agarose gel after electrophoresing at 100 V for 35 min, and stained with ethidium bromide (200 mL of 0.5 μg/μL) for 10 min. Finally, the results were imaged by an ultraviolet transilluminator (BioDoc-It™ Imaging System, Germany) with the exposure time of 1s.

2.11. Statistical analysis

All data were analyzed by calculating averages and standard deviations (SDs). No other statistical methods were used.

3. Results and discussion

3.1. Characterization of the micro-components

In order to optimize the operating conditions of the micro-components, the relationship between the pumping volume and applied gauge pressures was first investigated with PBS. The applied positive gauge pressure was 25 kPa while five different negative gauge pressures (−20, −30, −40, −50, & −60 kPa, respectively) were applied for the large (#1–4 and #6) and small (#5) micropumps (See Fig. 2b for details). As shown in Fig. 3 a–b, the optimal conditions for pumping ∼20 μL was an applied negative gauge pressure of −35 kPa for micropumps #1–4. Micropump #6 shows the similar trend. Micropump (#5) could only transport 3–4 μL at the similar condition, and variation in pumping volume never exceeded 1 μL (n = 5). Note that in order to ensure that the positive gauge pressure was the same for all other micropumps, only the driving frequency was changed.

Fig. 3.

(a) The pumping volume at a constant positive gauge pressure of 25 kPa and five different negative gauge pressure (−20, −30, −40, −50, & −60 kPa) for the micropumps (#1–4 and #6). They pumped ∼20 μL at a negative gauge pressure of −35 kPa (n = 5). (b) The micropump (#5) pumped ∼ 3–4 μL at a negative gauge pressure of −35 kPa (n = 5). (c) The mixing indices of micromixers 1 and 2 reach almost 100% within 2.5 s at a frequency of 1 or 2 Hz and a negative pressure of −35 kPa (n = 3).

The performance of the micromixers (#1 and #2, Fig. 2b) was also explored by measuring the correlation between the mixing index and the mixing time at different applied gauge pressures and driving efficiencies (1 vs. 2 Hz; n = 3) [41]. A fully homogenous mixture was obtained in only 3 s at a driving frequency of 2 Hz at a constant positive gauge pressure of 25 kPa and a constant negative gauge pressure of −35 kPa (Fig. 3c). However, because we aimed to mix biological samples with reagents, we nevertheless opted to use a 1 Hz driving efficiency in the subsequent experiments. It is worth noting that to the best of our knowledge, it is the first work to demonstrate the screening of aptamers targeting S1 proteins of SARS-CoV-2 in a fully automatic format. Moreover, each round of screening was performed automatically without manual intervention during the SELEX process. Furthermore, in order to reduce the human errors in SELEX operation, the micro-components on the pneumatic-based microfluidic chip could be optimally set up by users. For example, we set up a constant positive gauge pressure of 25 kPa and a constant negative gauge pressure of −35 kPa for the working conditions. As shown in Fig. 3a–b, the error bars of the pumping volume in the same pressure condition were relatively small (less than 5%). It means that the transport volume was stable in the same condition which could greatly reduce human errors.

Moreover, the newly developed microfluidic chip was equipped with a new long-type micropump #6 for transportation of ssDNA library/PCR product from the denaturation chamber (located in the heating region in Fig. 2b) towards the regeneration chamber (located in the cooling region in Fig. 2b) within a single-step pumping process, thus minimizing the dead volume and allowing efficient transportation of products while compared to previous works with two-step transport micropumps [42]. The new chip was also capable of maintaining stable temperatures over the chip between the two regions such that no pre-loaded reagents and samples would be damaged.

3.2. Screening results of on-chip SELEX

After each round of on-chip SELEX, an aliquot of amplified product was collected from the PCR chamber and electrophoresed on 2% agarose gels (Fig. 4 ). Correctly-sized, 72-bp bands could be clearly seen, with band intensity increasing with number of positive selection rounds (Fig. 4a). This confirms that ssDNA targeting the SARS-CoV-2 S1 protein was successfully screened in each round.

Fig. 4.

(a) Agarose gel electrophoresis of positive (S1-5; with SARS CoV-2) and competitive selection results (S6-7; with saliva). L = 50-bp DNA ladder. (b) Agarose gel electrophoresis of negative selection with influenza viruses A and B and one positive (S8). (c) Agarose gel electrophoresis of negative selection with four kinds of bacteria (S9): MRSA, Pseudomonas aeruginosa, Acinetobacter baumannii, and Klebsiella pneumoniae. (d) Agarose gel electrophoresis of negative selection with two cancer cells (S10): A549 (human lung cancer cell line) and DOK (human dysplastic oral keratinocyte cell line). (e) The 2-D structure of the screened aptamer for SARS-CoV-2 S1 (J15).

After two rounds of saliva-competitive selection and three rounds of negative selection, the in vivo affinity and specificity were both theoretically improved; the aptamers with a high affinity towards SARS-CoV-2 in rounds 6 to 10 were also successfully amplified (Fig. 4b–d). Overall, 10 rounds of SELEX (including on-chip PCR) could be completed within 20 hr, which is significantly less than the one required by traditional SELEX processes.

3.3. Sequence, 2D structure, affinity, and specificity of the screened aptamer

After 10 rounds of screening, the final products were cloned and sequenced, and one aptamer, J15, comprised 30% of all sequences. It was then synthesized, and its 2D structure was shown in Fig. 4e. Note that sequences with lower free energy, complicated structures (more stem-loop structures), and more intra-molecular hydrogen bonds were considered to be desirable as probes.

To determine its Kd value, the aptamer was FAM-labelled, incubated with SARS-CoV-2 S1 protein-coated beads, and a value of 63.0 nM was measured based on the method we described previously (Fig. 5 a). It is noted that the Kd value was also determined by SPR as 93.8 nM (Fig. 5b–d). Note that the K_d measurement shown in Fig. 5a was derived by GraphPad Prism software (ver. 5.0, USA). Besides, Analysis of binding events (Anabel), Version 2.2.3, a free online software, was used to analyze the SPR data. Briefly, after loading the data acquired from SPR experiments into the Anabel server (Fig. 5b), the single drift correction and Y-axis adjustment were performed by selecting a baseline region of all curves. Next, a time zone ranging from 186 to 280 were zoomed for curve fitting (Fig. 5c). Then, the equilibrium RUs were derived from the fitting results, together with corresponding aptamer concentration plotted with Origin software (OriginLab, MA, US) and using Logistic model for deriving the Kd value (Fig. 5d).

Fig. 5.

(a) The K_d measurement of aptamer J15 from a FAM-based dilution series. Prism software determined the K_d to be 63.0 nM (n = 3) based on a curve-fitting algorithm. (b) The results of the SPR curves. (c) The fitted curves of designated aptamer concentrations given by the Anabel online data analysis. (d) The equilibrium relative units (RUs) versus corresponding aptamer concentrations were plotted and fitted with the Logistic model. The derived K_d value is 93.8 nM.

By incubating this aptamer with off-target bacteria, viruses and cells (Figs. S1–S3), it was clear that it is highly specific to SARS-CoV-2 S1 proteins. This high specificity is likely in part due to the three rounds of negative selection by using two non-target viruses (influenza A and B), two kinds of cancer cells (adenocarcinoma human alveolar basal epithelial cells and dysplastic oral keratinocytes), and four bacteria (methicillin-resistant Staphylococcus aureus, Pseudomonas aeruginosa, Acinetobacter baumannii and Klebsiella pneumoniae).

3.4. Aptamer binding with pseudo-viruses

Experimental results show that this aptamer (J15) exhibited high binding ability with different kinds of SARS-CoV-2 pseudo-viruses (Fig. 6 ). The ability to identify different spike proteins was therefore an advantage for the screened aptamer. Furthermore, it was also helpful

Fig. 6.

Fluorescent images of aptamer-coated beads alone or with different SARS-CoV-2 pseudo-viruses.

in the subsequent SARS-CoV-2 detection. Moreover, the variety of SARS-CoV-2 mutants is increasing. According to the recent reports, BA.4 and BA.5 could escape the antibodies produced by vaccines [43]. Among them, the probability of BA.4 and BA.5 viruses escaping from the antibodies produced by the vaccine was 4 times higher than that of BA.2. This also means that BA.4 and BA.5 are more likely to cause breakthrough infections, and individuals who have previously been infected may be re-infected. However, if different types of SARS-CoV-2 mutants could be detected by using the screened aptamer, it could be applied to various diagnostic and treatment options while it could increase the accuracy of detection. In addition, two kinds of pseudo-viruses including original (Wuhan) and BA.4/BA.5 (Omicron) strains have been captured by specific aptamer-coated beads in saliva samples (Fig. S4). This result demonstrated the effectiveness of competitive selection by saliva in SELEX and the bio-applicability of the selected aptamers for SARS-CoV-2 in saliva samples.

3.5. Aptamer binding with inactivated SARS-CoV-2 virus particles and clinical isolates of SARS-CoV-2 viruses

The gel electrophoretic results of SARS-CoV-2 S1 aptamer conjugated beads binding with inactivated viruses (100 μL, 10⁴ copies/mL) was shown in Fig. 7 a. According to the results, it could confirm that this aptamer exhibited satisfactory capability to bind with the inactivated SARS-CoV-2 virus particles.

Fig. 7.

(a) Agarose gel electrophoresis of inactivated virus binding results. In all panels, L = 100-bp DNA ladder and N = negative control (ddH2O). In (a), Lane 1 = positive control (purified SARS-CoV-2 RNA [10⁴ copies/mL]), Lane 2 = SARS-CoV-2 S1 aptamer-conjugated beads with ddH₂O (100 μL), and Lane 3 = SARS-CoV-2 S1 aptamer-conjugated beads with inactivated viruses (100 μL, 10⁴ copies/mL). (b) Limit of detection (LOD) tests with clinical isolate of SARS-CoV-2 viruses (Wuhan strain) by RT-PCR (E gene). Lane 1 = positive control (1.95 × 10¹⁰ copies/mL). Lanes 2–8 = SARS-CoV-2 S1 aptamer-conjugated beads with 1x, 10¹x, 10²x, 10³x, 10⁴x, 10⁵x, 10⁶ x, 10⁷ x and 10⁸ x-diluted samples, respectively. (c) LOD tests with clinical isolate of SARS-CoV-2 viruses (Omicron BA.5 strain) by RT-PCR (E gene). Lane 1 = positive control (1.95 × 10¹⁰ copies/mL). Lanes 2–8 = SARS-CoV-2 S1 aptamer-conjugated beads with 1 x, 10¹x, 10²x, 10³x, 10⁴x, 10⁵x, 10⁶x, 10⁷x and 10⁸x-diluted samples, respectively.

For binding the clinical isolates of SARS-CoV-2 viruses, the J15 aptamer bound both the Wuhan and Omicron BA.5 strains, and a 125-bp RT-PCR product band was clearly observed (Fig. 7b–c). The aptamer J15 which was selected from this automated microfluidic system was capable of capturing inactivated viruses, eight different pseudo-viruses (i.e. SARS-CoV-2 mutants), and clinical isolates of target viruses. It is worth noting that the limits of detection (LOD) (4.80×10¹ copies/mL and 1.95×10² copies/mL were determined for Wuhan and BA.5 strains, respectively) are comparable to the results from different commercial kits (around 2.50×10¹ copies/mL to 5.00×10² copies/mL [40,43,44]. It indicates that the proposed method using aptamers to capture viruses, then followed by RT-PCR could be useful for fast and accurate diagnosis in the near future. Moreover, the LOD of SARS-CoV-2 diagnosis also meets the standard recommended by the WHO RT-PCR assay [17]. This superior LOD for the former is likely due to the fact that the original SELEX featured S1 proteins of the Wuhan strain as targets; in contrast, BA.5 is a relatively newer mutant, though the J15 aptamer nevertheless demonstrated its capability to bind it. Although we used the Wuhan strain as the screening object, this aptamer exhibited a satisfactory capturing ability for various variants. It is speculated that the shape of the folding of the aptamer has a common binding site with the same S1 spike proteins of different SARS-CoV-2 viruses, thus revealing its usability for clinical applications. Moreover, it shows that the aptamer could be used for detecting multiple variants as long as S1 proteins exist. Furthermore, the microfluidic chip has advantages, such as automation and rapid screening when compared with conventional SELEX. If the developed aptamer could not detect new variants, the aptamer will be selected for the new virus by the developed microfluidic system. For discriminating multiplex variants using multiplex aptamer assays, multiple aptamers will be considered as the next step selection.

In addition, another advantage of this device is that it could be customized and adjusted by the users based on the suitable experimental conditions (i.e. number of cycles and type of natural selection). For instance, the entire screening process includes five rounds of positive selection (with the S1 protein), two rounds of competitive selection with human saliva and three rounds of negative selection to enhance the specificity of the screened aptamer. As shown in Figs. S1–S3, specificity of the screened aptamer has been greatly enhanced. Furthermore, from the results of Fig. 4, it confirms that ssDNA targeting the SARS-CoV-2 S1 spike protein was successfully screened in each round. With the two rounds of saliva-competitive selection and three rounds of negative selection, the in vivo affinity and specificity were both significantly improved. Moreover, 10 rounds of SELEX could be completed within 20 hr, and the average time for 1 round of SELEX is only 2 hr, which is significantly reduced when compared with conventional SELEX round (about days to months). Moreover, the newly developed microfluidic chip was equipped with a new long-type micropump #6 for transportation of ssDNA library/PCR product from the denaturation chamber (located in the heating region) towards the regeneration chamber (located in the cooling region) within a single-step pumping process, thus minimizing the dead volume and allowing efficient transportation of products while compared to previous works with two-step transport micropumps [42]. The new chip was also capable of maintaining stable temperatures over the chip between the two regions such that no pre-loaded reagents and samples would be damaged.

4. Conclusions

This work has demonstrated an integrated microfluidic platform equipped with micromixers, micropumps, microvalves and on-chip PCR to effectively automate the entire SELEX process for screening of SARS-CoV-2 S1 aptamer probes. Each round could be finished within 2 hr and overall five rounds of positive selection, two rounds of saliva-competitive selection, and three rounds of negative selection were carried out. The resulting aptamer, J15, was characterized by a Kd of 63.0 nM and could bind eight strains of pseudo-viruses as well as inactive viruses. LODs for the Wuhan and Omicron strains were satisfactory for clinical applications (4.80 × 10¹ and 1.95×10² copies/mL, respectively), signifying that the screened aptamer could be promising for diagnosis of SARS-CoV-2. The better LOD of the current study should attribute to molecular diagnosis using PCR, which provides a better LOD when compared with traditional antigen sensors. In addition to sensitivity, rapid diagnostic assay such as serological tests may also suffer from antibody cross-reactions with other proteins or pathogens to decrease their specificity. In this work, the aptamer captured SARS-COV-V2 viruses and amplified target genes by specific primers in the following RT-PCR step with the captured viral RNA. Therefore, the nucleic acid technology (NAT) had high specificity and sensitivity than antigen/antibody sensors.

CRediT authorship contribution statement

Hung-Bin Wu: Writing – original draft, Data curation, Investigation, Validation, Formal analysis, Methodology. Chih-Hung Wang: Writing – review & editing, Conceptualization, Methodology, Validation. Yi-Da Chung: SPR measurement. Yan-Shen Shan: Writing – review & editing, Conceptualization, Provided, Resources, Validation. Ying-Jun Lin: Data curation, Investigation, Validation. Huey-Pin Tsai: Writing – review & editing, Conceptualization, Methodology, Provided, Resources, Validation. Gwo-Bin Lee: Writing – original draft, Writing – review & editing, Conceptualization, Methodology, Provided, Resources, Validation, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.