Biosensors and Bioelectronics 165 (2020) 112356

Available online 3 June 20200956-5663/© 2020 Elsevier B.V. All rights reserved.

Homogeneous circle-to-circle amplification for real-time optomagnetic detection of SARS-CoV-2 RdRp coding sequence

Bo Tian a,*, Fei Gao b, Jeppe Fock c, Martin Dufva a, Mikkel Fougt Hansen a,**

a Department of Health Technology, Technical University of Denmark, DTU Health Tech, Building 345C, DK-2800, Kongens Lyngby, Denmark b Department of Physics, Technical University of Denmark, DTU Physics, Building 307, DK-2800, Kongens Lyngby, Denmark c Blusense Diagnostics ApS, Fruebjergvej 3, DK-2100, Copenhagen, Denmark

A R T I C L E I N F O

Keywords: Rolling circle amplification Magnetic nanoparticle Optomagnetic sensing SARS-CoV-2 COVID-19

A B S T R A C T

Circle-to-circle amplification (C2CA) is a specific and precise cascade nucleic acid amplification method con-sisting of more than one round of padlock probe ligation and rolling circle amplification (RCA). Although C2CA provides a high amplification efficiency with a negligible increase of false-positive risk, it contains several step- by-step operation processes. We herein demonstrate a homogeneous and isothermal nucleic acid quantification strategy based on C2CA and optomagnetic analysis of magnetic nanoparticle (MNP) assembly. The proposed homogeneous circle-to-circle amplification eliminates the need for additional monomerization and ligation steps after the first round of RCA, and combines two amplification rounds in a one-pot reaction. The second round of RCA produces amplicon coils that anneal to detection probes grafted onto MNPs, resulting in MNP assembly that can be detected in real-time using an optomagnetic sensor. The proposed methodology was applied for the detection of a synthetic complementary DNA of SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2, also known as 2019-nCoV) RdRp (RNA-dependent RNA polymerase) coding sequence, achieving a detection limit of 0.4 fM with a dynamic detection range of 3 orders of magnitude and a total assay time of ca. 100 min. A mathematical model was set up and validated to predict the assay performance. Moreover, the proposed method was specific to distinguish SARS-CoV and SARS-CoV-2 sequences with high similarity.

1. Introduction

Rolling circle amplification (RCA) is the most widely adopted isothermal nucleic acid amplification method due to its simplicity, ac-curacy, versatility, and robustness (Ali et al., 2014; Mohsen and Kool, 2016; Zhao et al., 2008). Combined with a target-mediated padlock probe (PLP) ligation step, RCA has an ultrahigh specificity for the distinction of even single-nucleotide variants (Ban�er et al., 1998; Nils-son, 2006; Nilsson et al., 1994; Qi, 2001). The ligation-RCA strategy has been applied on different sensors for different kinds of targets (Ji et al., 2012; Li et al., 2018; Zhang et al., 2018; Zhao et al., 2020). However, the linear amplification format limits the amplification efficiency of con-ventional ligation-RCA assays, resulting in typically picomolar-level limits of detection (LOD) for nucleic acid quantification. To improve the amplification efficiency, different RCA-based approaches were developed with exponential or quadratic amplification formats. RCA-based exponential reactions, such as hyperbranched-RCA (Lizardi

et al., 1998; Yong Zhang et al., 1998) and primer generation-RCA (Murakami et al., 2012, 2009), recycle amplicons as primers or tem-plates, achieving attomolar to sub-femtomolar level LODs at the expense of false-positive results caused by off-template polymerase products, contaminants, and side-reactions (Tan et al., 2008; Tian et al., 2020b; Zhao et al., 2015). RCA-based cascade amplification strategies combine amplification reactions tandemly, thus improving the efficiency without increasing the risk of false-positive results.

Circle-to-circle amplification (C2CA) is a typical RCA-based cascade amplification that has been applied for nucleic acid analysis and single molecule detection (Dahl et al., 2004; Ke et al., 2011; Kühnemund et al., 2014). In C2CA, amplicons of the first round RCA are converted into multiple circles by monomerization (endonuclease digestion) and liga-tion. Newly formed circles are then employed as templates for the sec-ond round RCA. A two-round C2CA can achieve a femtomolar LOD (Mezger et al., 2015; Wei et al., 2016), and additional rounds of amplification can be performed by repeating the monomerization,

* Corresponding author. ** Corresponding author.

E-mail addresses: botia@dtu.dk (B. Tian), mfha@dtu.dk (M.F. Hansen).

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

journal homepage: http://www.elsevier.com/locate/bios

https://doi.org/10.1016/j.bios.2020.112356 Received 24 March 2020; Received in revised form 28 May 2020; Accepted 1 June 2020

Biosensors and Bioelectronics 165 (2020) 112356

2

ligation and RCA processes. However, C2CA can only be performed step-by-step because: (a) primers of the second round RCA can change the first round RCA to hyperbranched-RCA; and (b) monomerization and ligation are incompatible processes. Therefore, C2CA requires several operation steps, different reaction temperatures (heat inactivation is needed after monomerization), and a complex set of DNA primer-s/templates/triggers, which limit its utilization in biosensing.

Herein, a novel homogeneous and isothermal C2CA strategy is re-ported for nucleic acid quantification, which simplifies the operation process and combines two RCA rounds in a one-pot amplification re-action with simultaneous real-time detection. The proposed homoge-neous circle-to-circle amplification (HC2CA) strategy consists of a PLP ligation step followed by a homogeneous amplification and detection step. Intermediate amplicons can be directly used as primers for next round RCA, eliminating the need for additional monomerization and ligation. End amplicons of HC2CA, i.e., single-stranded DNA (ssDNA) coils, hybridize with detection probes grafted onto magnetic nano-particles (MNPs), leading to the assembly of MNPs. Individual MNPs and MNP assemblies have different properties of light scattering and ab-sorption, which vary with the rotation of these magnetic objects in response to an applied oscillating magnetic field. Based on this effect, an optomagnetic sensor was utilized to analyze the state of MNPs in the HC2CA suspension, achieving a real-time volumetric sensing format.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the etiological agent of Coronavirus Disease 2019 (COVID-19) starting outbreak in late December 2019, has rapidly become a major concern throughout the world (Chan et al., 2020b; Zhou et al., 2020; Zhu et al., 2020). Rapid diagnosis of SARS-CoV-2 is essential for case identification, contact tracing and evaluation of infection control (Chan et al., 2020a, 2015; Peiris et al., 2003). Thanks to the early availability of SARS-CoV-2 complete genome, standardized laboratory protocols of real-time RT-PCR assays were published online on January 13, 2020 targeting RNA-dependent RNA polymerase (RdRp), envelope, and nucleocapsid genes of SARS-CoV-2 (Corman et al., 2020). Among these three assays, the RdRp-based assay showed a decent performance evaluated using synthetic nucleic acids, and has been widely implemented in Europe (Reusken et al., 2020). In this study, we selected a highly conserved region of the SARS-CoV-2 RdRp coding gene, and the synthetic com-plementary DNA (cDNA) of the selected sequence was utilized for the design and validation of the HC2CA-based optomagnetic biosensor.

2. Materials and methods

2.1. Chemicals and DNA sequences

Circular templates for the first round RCA (CT1) were prepared by a PLP-ligation process using Ampligase, and the reagents were obtained from Nordic Biolabs (T€aby, Sweden). Circular templates for the second round RCA (CT2) were prepared by a different ligation process using CircLigase II, and the reagents were obtained from Biosearch Technol-ogies (Novato, CA, USA). Phi29 polymerase, phi29 buffer, bovine serum albumin (BSA), SYBR Gold, Tris-HCl buffer (1 M, pH 8.0), and Tris- acetate-EDTA buffer (TAE, 50�) were purchased from Thermo Fisher Scientific (Waltham, MA, USA). AluI (restriction endonuclease), ther-molabile exonuclease I, exonuclease III, and dNTPs were purchased from New England BioLabs (Ipswich, MA, USA). Fetal bovine serum (FBS) was obtained from Sigma-Aldrich (St. Louis, MO, USA). Streptavidin-coated iron oxide particles (100 nm size MNP, product code 10-19-102) were supplied by Micromod Partikeltechnologie GmbH (Rostock, Germany). Synthetic DNA sequences (listed in Tables S1 and S2) were prepared by Integrated DNA Technologies (Coralville, IA, USA). Unless otherwise specified, the buffer solution used in this study was Tris-HCl (50 mM, pH 8.0).

2.2. Functionalization of magnetic nanoparticles

MNPs (80 μL, 10 mg/mL) were washed twice and resuspended to 1 mg/mL before functionalization. The biotinylated detection probe was added to the MNP suspension to a concentration of 0.25 μM. Thereafter the suspension was incubated at 37 �C for 30 min, washed 3 times, and resuspended to a nanoparticle concentration of 1 mg/mL. Functional-ized MNPs were stored at 4 �C prior to use. Dynamic light scattering and scanning electron microscopic characterizations of ssDNA- functionalized MNPs were provided in our previous work using iden-tical functionalization method and MNPs (Tian et al., 2018a).

2.3. Preparation of circular templates

CT2 were prepared by a CircLigase II ssDNA ligation process. The ligation mixture consisted of linear CT2 sequences (0.5 μM), CircLigase II reaction buffer (1�), MnCl2 (2.5 mM), betaine (1 M), and CircLigase II ssDNA ligase (5 U/μL). The ligation reaction was conducted at 60 �C for 3 h, followed by 15 min of heat inactivation at 80 �C. Exonucleases I and III were added to the solution to final concentrations of 1 U/μL and 5 U/ μL, respectively. The digestion was conducted at 37 �C for 2 h, followed by 30 min of heat inactivation at 80 �C.

2.4. Padlock probe ligation, HC2CA, and conventional RCA

CT1 were prepared by a PLP ligation process. A ligation mixture for 20 nM of CT1 was composed of AmpL buffer (1�), BSA (0.2 μg/μL), PLP (60 nM), target DNA (20 nM), and Ampligase (0.25 U/μL). The ligation reaction was conducted at 55 �C for 10 min. Thereafter, the solution was diluted with buffer to obtain different concentrations of CT1 (hybridized with target DNA). For HC2CA, the reaction suspension consisted of phi29 buffer (2�), BSA (0.4 μg/μL), dNTPs (0.44 mM), CT2 (200 pM), functionalized MNPs (0.1 mg/mL), AluI (0.44 U/μL), and phi29 (0.67 U/ μL). The HC2CA suspension was mixed with (diluted) ligation mixture in a volumetric ratio of 1:1, followed by on-chip incubation at 37 �C for amplification and detection. The PLP ligation process for the conven-tional RCA was performed in an identical protocol as described above but using linear CT2 sequence and CT2 primer instead of PLP and target DNA, respectively. Except for the absence of CT2 and AluI, conventional RCA were performed using the same procedure as HC2CA.

2.5. Optomagnetic measurement

A detailed description of the optomagnetic technique is provided in Supplementary Section S1. After loading the amplification suspensions, optomagnetic detection chips (each containing 90 μL of the reaction suspension) were sealed and mounted in the optomagnetic setup. Before each optomagnetic measurement, a magnetic incubation process (10 cycles each containing 2 s of 2.6 mT field followed by 2 s of 0 mT) was automatically applied. Each optomagnetic spectrum consisting of 41 logarithmically equidistant frequencies was recorded in about 1 min.

2.6. Gel electrophoresis

Agarose gel electrophoresis (2.5%, 1� TAE buffer) was applied to analyze amplicons of different amplification reactions. Samples were prepared by mixing amplicons with loading buffer and SYBR Gold, and analyzed at room temperature by 30 min of electrophoresis performed at 100 V. Gel images were taken by a UVP BioSpectrum Imaging System (Analytik Jena, Jena, Germany).

3. Results and discussion

3.1. Amplification principle

A 50-nt-long highly conserved region of SARS-CoV-2 RdRp coding

B. Tian et al.

Biosensors and Bioelectronics 165 (2020) 112356

3

sequence was selected, and its cDNA was synthesized as the target sequence (Table S1). For the detection of ssDNA target, the assay is initiated with a target-mediated PLP ligation step to form CT1 (circular template for the first round RCA) with primers. HC2CA consists of two rounds of RCA. First, a nicking-enhanced RCA (NickRCA) produces single-stranded intermediate amplicons (Fig. 1, blue arrows) (Gao et al., 2018; Huang et al., 2017; Li et al., 2010; Xiang et al., 2015). This is followed by a second round of RCA that utilizes intermediate amplicons as primers on an added circular template (CT2) to produce amplicon coils (Fig. 1, red arrows). The PLP for NickRCA is designed with a cutting site (nucleotide sequence of AGCT) protected by the 5-methylated deoxycytidine, resulting in nicking but not cutting when reacted with the restriction endonuclease AluI. Due to the nicking reaction that continuously provides nicks on the amplicon strand, polymerases can act tandemly on the same template (Tian et al., 2019a). End amplicons are long ssDNA molecules that form coils in the suspension. These coils contain periodic sequences that can hybridize with the detection probe on the MNPs, leading to the assembly of MNPs (Akhtar et al., 2010; Tian et al., 2018b; Zard�an G�omez de la Torre et al., 2010). The formation and size increase of MNP assemblies can be monitored in real-time by an optomagnetic sensor (Supplementary Section S1 and Fig. S1) (Fock et al., 2018, 2017b; 2017a; Tian et al., 2019b). Utilization of MNPs also enables magnetic incubation during the assay, which can dramatically increase the reaction rate and promotes the assay towards point-of-care applications (Baudry et al., 2006; Dayn�es et al., 2015; Xianyu et al., 2018).

3.2. Feasibility test

In our previous study on NickRCA, a 3’-end blocked short ssDNA was added to the NickRCA suspension to serve as (a) an amplicon protector that annealed to the 3’-end of the amplicon to prevent the exonucleo-lytic digestion by phi29, and (b) a restriction oligo that annealed to the cutting sites in amplicon dimers/concatemers and induced cutting to produce amplicon monomers (Tian et al., 2019a). In the HC2CA design, intermediate amplicons should have 3’-ends that can be amplified or trimmed by phi29 after hybridized onto CT2. Therefore, using a re-striction oligo can theoretically provide more intermediate amplicons (by breaking dimers/concatemers into monomers), but influences the hybridization between the intermediate amplicon and CT2. Gel elec-trophoresis was applied to evaluate the influence of the restriction oligo by amplifying 10 pM of target DNA by NickRCA for 60 min with or without the restriction oligo. As shown in Fig. 2a, NickRCA without

restriction oligo (lane 1) contains no observable amplicon

Fig. 1. Schematic illustration of homogeneous circle-to-circle amplification. In the first round of RCA, polymerases act tandemly to generate intermediate amplicons. Intermediate amplicons anneal to CT2 for the second round of RCA, generating amplicon coils that lead to the assembly of MNPs. After a ligation step, all processes of amplification, hybridization, and detection take place simultaneously on-chip at 37 �C.

Fig. 2. Feasibility test for HC2CA. (a) Agarose gel electrophoresis analysis for the amplicons of 10 pM target NickRCA (lane 1), 10 pM target NickRCA with 10 nM of restriction oligo (lane 2), and 100 fM target HC2CA with 100 pM of CT2 (lane 3). (b) End-point (90 min of reaction) ϕ-spectra of NickRCA (100 fM of target, blue curve), HC2CA (100 fM of target, 100 pM of CT2, red curve), and blank control of HC2CA (black curve). The gray zone in panel (b) indicates the frequency range for the ϕ value calculation. (For interpretation of the refer-ences to colour in this figure legend, the reader is referred to the Web version of this article.)

B. Tian et al.

Biosensors and Bioelectronics 165 (2020) 112356

4

dimers/concatemers (which would produce a ladder pattern). The band of amplicons in lane 2 is slightly darker than the band in lane 1, due to the hybridization of 10 nM of 27-nt-long restriction oligo. For target concentrations lower than 10 pM, the ratio between nickase and nicking points will be even higher due to the less acting NickRCA reactions, which further reduces the chance of generating amplicon dimer-s/concatemers. Therefore, the utilization of restriction oligo is not necessary when we focus on femtomolar target concentrations. In the electrophoresis study, we also demonstrated the feasibility of HC2CA by amplifying 100 fM of target DNA, generating amplicon coils that stayed in the sample-loading well (Fig. 2a, lane 3).

We next compared the performance of NickRCA and HC2CA for the detection of 100 fM target DNA (the final concentration in the suspen-sion) using an optomagnetic setup. Both amplification reactions were performed at 37 �C for 90 min. The optomagnetic phase lag (ϕ) spectra of NickRCA, HC2CA, and blank control (for HC2CA) presented in Fig. 2b shows that H2CA provided a much stronger change of the optomagnetic ϕ-spectrum, suggesting a higher amplification efficiency. Compared to the spectrum of blank control, more negative ϕ-values at low fre-quencies indicate the existence of MNP assemblies. Herein, we use the mean ϕ-value in the interval 1–10 Hz (gray zone in Fig. 2b), denoted as ϕ, to represent the formation and hydrodynamic size increase of MNP assembly. The ϕ-spectrum of conventional RCA-based detection of 100 fM target was indistinguishable from the blank control spectrum and is thus not shown.

3.3. Optimization of circular template concentration

CT2 used for the second round of RCA influences the HC2CA per-formance mainly in two aspects: it determines the maximum amplifi-cation efficiency; it competes with detection probes for hybridization with amplicon coils. Considering that CT2 matches the amplicon coil perfectly, we can infer that (a) a proportion of CT2 (depending on the ratio between concentrations of CT2 and target) hybridizes with the amplicon coils, and thus cannot anneal to the intermediate amplicons for the second round of RCA; and (b) formation of MNP assembly can be suppressed by the unused CT2 (CT2 not hybridized with any kind of amplicons) due to the competitive binding. Then a further assumption can be made that a higher CT2 concentration will result in a longer dead time during which no signal can be observed, but a higher rate of signal increase after the dead time. A simple mathematical model was set up to predict the signal trend at different ratios between CT2 and target

concentrations (Supplementary Section S2 and Fig. S2). We evaluated different CT2 concentrations (Fig. S3) for the balance between the amplification efficiency and the total assay time, and chose a CT2 con-centration of 100 pM for the subsequent study.

3.4. HC2CA-based DNA quantification

Representative HC2CA time-resolved ϕ-spectra of different target concentrations (Fig. 3) reflected the hydrodynamic size changes of MNPs during the 90 min amplification reaction. The ϕ-spectra changed slowly at the beginning of measurement, which was caused by the competition between CT2 and detection probes (dead time). Due to the continuous generation of amplicon coils and depletion of CT2, MNP assemblies formed and grew after the dead time, which was reflected by ϕ-values at low frequencies becoming more negative. Nonspecific MNP assembly was not observed in the blank control spectra.

The changes of the real-time optomagnetic signals ϕ (i.e., � Δϕ) with respect to that measured after 3 min are shown vs. time in Fig. 4a. Re-sults of the first two measurements were discarded to exclude the in-fluence of temperature settling after chip mounting. As predicted by the mathematical model (Fig. S2), the slope of the curve (after the dead time) shows a monotonic positive correlation with the target concen-tration; whereas the dead time shows a monotonic negative correlation with the target concentration. Signal plateaus were reached but with different values, suggesting that the plateau was not caused by the depletion of MNPs. As described in our assumption, the dead time was caused by the competition between CT2 and detection probes, and the signal increased after the dead time indicated the depletion of CT2. Therefore, no new second round RCA reaction could be initiated after the dead time, and the signal increase depended only on the extension of existing amplicon coils. However, with the size increase of amplicon coil-induced MNP assemblies, (a) attachment of more MNPs to existing MNP assemblies provided diminishing hydrodynamic size contributions, and (b) the ϕ-valley shifted to lower frequencies beyond our detection range. As a result, signal plateaus of different values were observed.

End-point � Δϕ values obtained after 90 min of amplification were utilized for the target quantification. The proposed HC2CA-based detection showed a linear dose-response at the range of 0.1–10 fM target (Fig. 4b), with an LOD of 0.4 fM obtained based on the 3σ crite-rion. The cutoff value was the average � Δϕ value of blank controls plus three standard deviations (σ). The whole dose-response curve (Fig. 4c,

Fig. 3. Representative time-resolved ϕ-changes during 90 min of HC2CA. The blue curves in each panel represent the initial spectra at 1 min, and the red curves represent the final spectra at 90 min. Target concentrations are indicated in each panel. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

B. Tian et al.

Biosensors and Bioelectronics 165 (2020) 112356

5

red circles) had a dynamic detection range of approximately 3 orders of magnitude, with no hook effect observed. For comparison, the LOD obtained by the single-round RCA based DNA quantification (Fig. 4c, black squares) was 300 fM. The cutoff value for the single-round RCA based strategy was not plotted due to overlap of the two cutoff lines. For all measured points in the dynamic detection range, the average co-efficients of variation were 6.91% for HC2CA and 6.75% for single- round RCA, suggesting that performing an additional round of RCA in a homogeneous strategy did not influence the repeatability significantly.

Generally, lower LODs can be achieved by prolonged reaction time in linear and quadratic amplification based assays. However, the dead time has to be considered in HC2CA. We recorded the dead time by using a time threshold for the signal reaching the cutoff value, and plotted the curve of dead time vs. target concentration (Fig. S4). The fitting curve (y ¼ 69:6x� 0:486) validated of our mathematical model (y∝ x� 0:5, Sup-plementary Section S2). Fig. S4 can be regarded as the dose-response curve for the real-time optomagnetic sensing format, and could be used to predict the time-dependent LOD for the end-point sensing format.

With an amplification time of 90 min, the HC2CA-based assay ach-ieved an LOD that was nearly 3 orders of magnitude lower than the LOD of the single-round RCA based assay. It was reported that highly proc-essive phi29 polymerase could copy a 100-nt-long circular template into a DNA strand containing approximately 1000 repeated complements in 60 min (Ban�er et al., 1998; Blanco et al., 1989), which implies that the two rounds of RCA reactions in HC2CA matched well in cascade form without significant loss in amplification efficiency. Considering that the NickRCA strategy is about 20 times more sensitive than conventional RCA (Tian et al., 2019a) and that the two rounds of RCA reactions took place simultaneously but not step-by-step, the proposed HC2CA should be more sensitive than the C2CA performed with the same amplification time. More importantly, C2CA contains several labor-intensive and time-consuming operation steps with different operation temperatures including ligation, amplification, polymerase inactivation, monomer-ization, endonuclease inactivation, second round ligation, and second round amplification; while HC2CA can be performed in one-pot after the ligation step and monitored in real-time. For comparison, results of several two-round C2CA based assays are given here. Mahmoudian et al. used C2CA with on-chip electrophoretic analysis for a target DNA of V. cholera, and achieved an LOD of 5 pM (12.5 ng/μL genomic DNA) (Mahmoudian et al., 2008). Pavankumar et al. applied C2CA with a lateral flow system for a target DNA of M. tuberculosis, and achieved an LOD of 0.2 pM (0.55 ng/μL genomic DNA) (Pavankumar et al., 2016). By combining C2CA with an optomagnetic sensor, Mezger et al. demon-strated an LOD of 105 CFU/mL E. coli, corresponding to a molar

concentration of 0.17 fM that is close to the LOD of our HC2CA-optomagnetic system, but with a total assay time of 4 h (Mezger et al., 2015).

Although not reported for biosensing applications, C2CA can combine more than two rounds of RCA steps to further lower the detection limit (Dahl et al., 2004). In contrast, HC2CA has a limitation here. HC2CA utilizes multiple intermediate amplicons as primers to trigger multiple next round RCA reactions, implying that NickRCA is required except for the last round. However, a NickRCA performed with abundant corresponding circular templates means that the NickRCA amplicons will preferentially anneal to the corresponding circular tem-plates but not the circular templates of next RCA round, resulting in a primer-generation RCA (Murakami et al., 2009), i.e., an exponential amplification format that is prone to give false-positive results. There-fore, HC2CA contains two RCA rounds at most. Some representative biosensors based on different RCA strategies are listed in Table S3 for further comparison.

3.5. Quantitative detection in serum

Although serological tests are not relevant in nucleic acid detection of SARS-CoV-2, we utilized FBS to evaluate the robustness of the HC2CA reaction as well as the MNP assembly based sensing system. Series di-lutions of target DNA were prepared in buffer solution containing 10% FBS. The final FBS concentration in the HC2CA suspension was 5%. Fig. 5a shows real-time optomagnetic signal changes when detecting 10% FBS samples. Due to matrix effects, for the same target concen-tration, the rate of signal increase found in FBS samples was lower than that obtained in pure buffer detection (cf. slopes in Figs. 4a and 5a). A strong hook effect was observed for the target concentration of 1 pM. Fortunately, the real-time � Δϕ spectra affected by hook effect could be identified according to the short dead time, suggesting that the shape of real-time spectra can be considered in addition to the end-point values. The average signal of blank controls (in 10% FBS) plus 3σ was plotted as the cutoff value (Fig. 5b). Due to the lowered reaction rates, HC2CA- based DNA quantification in FBS samples achieved an LOD of 1 fM, which was slightly higher than the LOD obtained in pristine conditions.

For the clinical detection of SARS-CoV-2, standard sample pretreat-ment steps of RNA extraction and cDNA preparation are required in practice. This means that the reaction suspension of ligation and HC2CA should have a roughly clean background with limited matrix effects. Therefore, the hook effect can be ignored unless detecting highly concentrated targets. Since HC2CA can be performed in one-pot without any additional operation steps, it can replace RCA and C2CA in other applications and be detected using conventional end-point readouts

Fig. 4. HC2CA-based DNA quantification in buffer. (a) Time-resolved ϕ (1–10 Hz) changes for different target concentrations amplified by HC2CA. (b) The linear dose-response range of HC2CA (90 min). (c) Dose-response curves of target detections based on HC2CA (red circles) and conventional RCA (black squares). Red dotted lines indicate the cutoff value. Error bars indicate the standard deviations of independent replicates (n ¼ 3). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

B. Tian et al.

Biosensors and Bioelectronics 165 (2020) 112356

6

such as a flow cytometer (for coil counting).

3.6. Specificity test

The PLP ligation step provides a strong specificity that can distin-guish single-nucleotide variants (Heo et al., 2016; Lizardi et al., 1998; Xiang et al., 2013). However, for non-target molecules that can anneal to PLP but cannot mediate the ligation reaction, side-reactions can be triggered. As illustrated in Fig. 6a, such non-target molecules can hy-bridize with linear PLP and be extended by polymerases, generating double-stranded nicking sequences. After the nicking reaction, the nick can be recognized and extended again by the polymerase, resulting in a strand displacement amplification (SDA) reaction (Tian et al., 2020a). However, in our molecular design, the SDA amplicons do not contain the CT2 binding segment (the gray sequence in Fig. 6a), and therefore cannot provide an optomagnetic signal.

Four types of synthetic nonspecific DNA sequences (Table S2) were measured at a concentration of 100 fM to evaluate the specificity of the proposed biosensor. Note that the selected SARS-CoV RdRp cDNA has only four nucleotides mismatch with the PLP. Due to the 24-nt-long perfect match between the non-target SARS-CoV sequence and the 3’- end of the PLP, the SDA side-reaction can be triggered during the in-cubation. Fig. 6b shows end-point � Δϕ values of specific and nonspe-cific targets obtained after 90 min of HC2CA. All nonspecific signals were lower than the cutoff value, indicating a high specificity of the proposed method.

4. Conclusions

In summary, we demonstrated a homogeneous and isothermal C2CA strategy and applied it for real-time optomagnetic DNA detection.

Compared to previously reported C2CA-based sensors, the proposed approach achieved a sub-femtomolar level detection limit and signifi-cantly simplified the operation by eliminating the labor-intensive and time-consuming operation steps requiring different reaction tempera-tures. Capability of target quantification in 10% FBS samples was demonstrated with an acceptable loss of sensitivity. Moreover, we dis-cussed the influence of side-reactions that could be caused by nonspe-cific molecules, and demonstrated the specificity of the proposed methodology to distinguish SARS-CoV and SARS-CoV-2 sequences with high similarity.

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.

CRediT authorship contribution statement

Bo Tian: Conceptualization, Methodology, Investigation, Visualiza-tion, Writing - original draft. Fei Gao: Methodology, Formal analysis, Visualization. Jeppe Fock: Software, Formal analysis. Martin Dufva: Conceptualization, Writing - review & editing. Mikkel Fougt Hansen: Conceptualization, Validation, Writing - review & editing.

Acknowledgments

This work was financially supported by H2020 Marie Skłodowska- Curie Actions (Grant No. 713683, COFUNDfellowsDTU), MUDP (Grant No. MST-141-01415) and DFF (Grant No. 4184-00121B).

Fig. 5. HC2CA-based DNA quantification in 10% FBS. (a) Time-resolved ϕ (1–10 Hz) changes of HC2CA for different target concentrations. (b) Dose- response curve. Error bars indicate the standard deviations of independent replicates (n ¼ 3).

Fig. 6. Specificity test. (a) Illustration of side-reactions caused by non-target molecules and unsuccessful ligation. Non-target molecules can anneal to the padlock probe and trigger a strand displacement amplification. (b) Changes of ϕ (1–10 Hz) after 90 min of HC2CA for 100 fM of indicated sequences. The red dotted line indicates the cutoff value. Error bars indicate the standard de-viations of independent replicates (n ¼ 3). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

B. Tian et al.

Biosensors and Bioelectronics 165 (2020) 112356

7

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.bios.2020.112356.

References

Akhtar, S., Str€omberg, M., Zard�an G�omez de la Torre, T., Russell, C., Gunnarsson, K., Nilsson, M., Svedlindh, P., Strømme, M., Leifer, K., 2010. J. Phys. Chem. B 114, 13255–13262.

Ali, M.M., Li, F., Zhang, Z., Zhang, K., Kang, D.K., Ankrum, J.A., Le, X.C., Zhao, W., 2014. Chem. Soc. Rev. 43, 3324–3341.

Ban�er, J., Nilsson, M., Mendel-Hartvig, M., Landegren, U., 1998. Nucleic Acids Res. 26, 5073–5078.

Baudry, J., Rouzeau, C., Goubault, C., Robic, C., Cohen-Tannoudji, L., Koenig, A., Bertrand, E., Bibette, J., 2006. Proc. Natl. Acad. Sci. U.S.A. 103, 16076–16078.

Blanco, L., Bernad, A., Lazaro, J.M., Martin, G., Garmendia, C., Salas, M., 1989. J. Biol. Chem. 264, 8935–8940.

Chan, J.F.W., Lau, S.K.P., To, K.K.W., Cheng, V.C.C., Woo, P.C.Y., Yuen, K.Y., 2015. Clin. Microbiol. Rev. 28, 465–522.

Chan, J.F.W., Yip, C.C.Y., To, K.K.W., Tang, T.H.C., Wong, S.C.Y., Leung, K.H., Fung, A.Y. F., Ng, A.C.K., Zou, Z., Tsoi, H.W., Choi, G.K.Y., Tam, A.R., Cheng, V.C.C., Chan, K. H., Tsang, O.T.Y., Yuen, K.Y., 2020a. J. Clin. Microbiol. https://doi.org/10.1128/ JCM.00310-20.

Chan, J.F.W., Yuan, S., Kok, K.H., To, K.K.W., Chu, H., Yang, J., Xing, F., Liu, J., Yip, C.C. Y., Poon, R.W.S., Tsoi, H.W., Lo, S.K.F., Chan, K.H., Poon, V.K.M., Chan, W.M., Ip, J. D., Cai, J.P., Cheng, V.C.C., Chen, H., Hui, C.K.M., Yuen, K.Y., 2020b. Lancet 395, 514–523.

Corman, V.M., Landt, O., Kaiser, M., Molenkamp, R., Meijer, A., Chu, D.K., Bleicker, T., Brünink, S., Schneider, J., Schmidt, M.L., Mulders, D.G., Haagmans, B.L., van der Veer, B., van den Brink, S., Wijsman, L., Goderski, G., Romette, J.L., Ellis, J., Zambon, M., Peiris, M., Goossens, H., Reusken, C., Koopmans, M.P., Drosten, C., 2020. Euro Surveill. 25, 2000045.

Dahl, F., Ban�er, J., Gullberg, M., Mendel-Hartvig, M., Landegren, U., Nilsson, M., 2004. Proc. Natl. Acad. Sci. U.S.A. 101, 4548–4553.

Dayn�es, A., Temurok, N., Gineys, J.P., Cauet, G., Nerin, P., Baudry, J., Bibette, J., 2015. Anal. Chem. 87, 7583–7587.

Fock, J., Balceris, C., Costo, R., Zeng, L., Ludwig, F., Hansen, M.F., 2018. Nanoscale 10, 2052–2066.

Fock, J., Jonasson, C., Johansson, C., Hansen, M.F., 2017a. Phys. Chem. Chem. Phys. 19, 8802–8814.

Fock, J., Parmvi, M., Str€omberg, M., Svedlindh, P., Donolato, M., Hansen, M.F., 2017b. Biosens. Bioelectron. 88, 94–100.

Gao, Z., Wu, C., Lv, S., Wang, C., Zhang, N., Xiao, S., Han, Y., Xu, H., Zhang, Y., Li, F., Lyu, J., Shen, Z., 2018. Anal. Bioanal. Chem. 410, 6819–6826.

Heo, H.Y., Chung, S., Kim, Y.T., Kim, D.H., Seo, T.S., 2016. Biosens. Bioelectron. 78, 140–146.

Huang, J., Li, X.Y., Du, Y.C., Zhang, L.N., Liu, K.K., Zhu, L.N., Kong, D.M., 2017. Biosens. Bioelectron. 91, 417–423.

Ji, H., Yan, F., Lei, J., Ju, H., 2012. Anal. Chem. 84, 7166–7171. Ke, R., Zorzet, A., G€oransson, J., Lindegren, G., Sharifi-Mood, B., Chinikar, S.,

Mardani, M., Mirazimi, A., Nilsson, M., 2011. J. Clin. Microbiol. 49, 4279–4285. Kühnemund, M., Witters, D., Nilsson, M., Lammertyn, J., 2014. Lab Chip 14, 2983–2992. Li, B., Yin, H., Zhou, Y., Wang, M., Wang, J., Ai, S., 2018. Sensor. Actuator. B Chem. 255,

1744–1752.

Li, J., Deng, T., Chu, X., Yang, R., Jiang, J., Shen, G., Yu, R., 2010. Anal. Chem. 82, 2811–2816.

Lizardi, P.M., Huang, X., Zhu, Z., Bray-Ward, P., Thomas, D.C., Ward, D.C., 1998. Nat. Genet. 19, 225–232.

Mahmoudian, L., Kaji, N., Tokeshi, M., Nilsson, M., Baba, Y., 2008. Anal. Chem. 80, 2483–2490.

Mezger, A., Fock, J., Antunes, P., Østerberg, F.W., Boisen, A., Nilsson, M., Hansen, M.F., Ahlford, A., Donolato, M., 2015. ACS Nano 9, 7374–7382.

Mohsen, M.G., Kool, E.T., 2016. Acc. Chem. Res. 49, 2540–2550. Murakami, T., Sumaoka, J., Komiyama, M., 2012. Nucleic Acids Res. 40, e22. Murakami, T., Sumaoka, J., Komiyama, M., 2009. Nucleic Acids Res. 37, e19. Nilsson, M., 2006. Histochem. Cell Biol. 126, 159–164. Nilsson, M., Malmgren, H., Samiotaki, M., Kwiatkowski, M., Chowdhary, B.P.,

Landegren, U., 1994. Science 265, 2085–2088. Pavankumar, A.R., Engstr€om, A., Liu, J., Herthnek, D., Nilsson, M., 2016. Anal. Chem.

88, 4277–4284. Peiris, J.S.M., Lai, S.T., Poon, L.L.M., Guan, Y., Yam, L.Y.C., Lim, W., Nicholls, J., Yee, W.

K.S., Yan, W.W., Cheung, M.T., Cheng, V.C.C., Chan, K.H., Tsang, D.N.C., Yung, R.W. H., Ng, T.K., Yuen, K.Y., 2003. Lancet 361, 1319–1325.

Qi, X., 2001. Nucleic Acids Res. 29, e116. Reusken, C.B.E.M., Broberg, E.K., Haagmans, B., Meijer, A., Corman, V.M., Papa, A.,

Charrel, R., Drosten, C., Koopmans, M., Leitmeyer, K., On behalf of evd-LabNet, 2020. Euro Surveill.. 25, 2000082.

Tan, E., Erwin, B., Dames, S., Ferguson, T., Buechel, M., Irvine, B., Voelkerding, K., Niemz, A., 2008. Biochemistry 47, 9987–9999.

Tian, B., Fock, J., Minero, G.A.S., Garbarino, F., Hansen, M.F., 2019a. Anal. Chem. 91, 10102–10109.

Tian, B., Fock, J., Minero, G.A.S., Hansen, M.F., 2020a. Biosens. Bioelectron. 160, 112219.

Tian, B., Han, Y., Fock, J., Str€omberg, M., Leifer, K., Hansen, M.F., 2019b. ACS Appl. Nano Mater. 2, 1683–1690.

Tian, B., Han, Y., Wetterskog, E., Donolato, M., Hansen, M.F., Svedlindh, P., Str€omberg, M., 2018a. ACS Sens. 3, 1884–1891.

Tian, B., Liao, X., Svedlindh, P., Str€omberg, M., Wetterskog, E., 2018b. ACS Sens. 3, 1093–1101.

Tian, B., Minero, G.A.S., Fock, J., Dufva, M., Hansen, M.F., 2020b. Nucleic Acids Res. 48, e30.

Wei, H., Tang, S., Hu, T., Zhao, G., Guan, Y., 2016. Sci. Rep. 6, 29229. Xiang, Y., Deng, K., Xia, H., Yao, C., Chen, Q., Zhang, L., Liu, Z., Fu, W., 2013. Biosens.

Bioelectron. 49, 442–449. Xiang, Y., Zhu, X., Huang, Q., Zheng, J., Fu, W., 2015. Biosens. Bioelectron. 66, 512–519. Xianyu, Y., Wang, Q., Chen, Y., 2018. Trac. Trends Anal. Chem. 106, 213–224. Yong Zhang, D., Brandwein, M., Chun Hung Hsuih, T., Li, H., 1998. Gene 211, 277–285. Zard�an G�omez de la Torre, T., Str€omberg, M., Russell, C., G€oransson, J., Nilsson, M.,

Svedlindh, P., Strømme, M., 2010. J. Phys. Chem. B 114, 3707–3713. Zhang, K., Lv, S., Lin, Z., Li, M., Tang, D., 2018. Biosens. Bioelectron. 101, 159–166. Zhao, W., Ali, M.M., Brook, M.A., Li, Y., 2008. Angew. Chem. Int. Ed. 47, 6330–6337. Zhao, X., Luo, C., Mei, Q., Zhang, H., Zhang, W., Su, D., Fu, W., Luo, Y., 2020. Anal.

Chem. 92, 5411–5418. Zhao, Y., Chen, F., Li, Q., Wang, L., Fan, C., 2015. Chem. Rev. 115, 12491–12545. Zhou, P., Yang, X.-L., Wang, X.-G., Hu, B., Zhang, L., Zhang, W., Si, H.-R., Zhu, Y., Li, B.,

Huang, C.-L., Chen, H.-D., Chen, J., Luo, Y., Guo, H., Jiang, R.-D., Liu, M.-Q., Chen, Y., Shen, X.-R., Wang, X., Zheng, X.-S., Zhao, K., Chen, Q.-J., Deng, F., Liu, L.- L., Yan, B., Zhan, F.-X., Wang, Y.-Y., Xiao, G.-F., Shi, Z.-L., 2020. Nature 579, 270–273.

Zhu, N., Zhang, D., Wang, W., Li, X., Yang, B., Song, J., Zhao, X., Huang, B., Shi, W., Lu, R., Niu, P., Zhan, F., Ma, X., Wang, D., Xu, W., Wu, G., Gao, G.F., Tan, W., 2020. N. Engl. J. Med. 382, 727–733.

B. Tian et al.