1

Article 1

Brief Communication: Magnetic Immuno-Detection of SARS-CoV-2 specific 2

Antibodies 3

Jan Pietschmann1,*, Nadja Vöpel1, Holger Spiegel1, Hans-Joachim Krause2 and Florian Schröper1 4

1 Fraunhofer Institute for Molecular Biology and Applied Ecology IME, Forckenbeckstraße 6, 52074 Aachen, 5

2 Institute of Biological Information Processing, Bioelectronics IBI-3, Forschungszentrum Jülich, 52428 Jülich, 6

Germany 7

* Corresponding author 8

E-mail: jan.pietschmann@ime.fraunhofer.de (JP) 9

Abstract: 10

SARS-CoV-2 causes ongoing infections worldwide, and identifying people with immunity is 11

becoming increasingly important. Available point-of-care diagnostic systems as lateral flow assays 12

have high potential for fast and easy on-site antibody testing but are lacking specificity, sensitivity 13

or possibility for quantitative measurements. Here, a new point-of-care approach for SARS-CoV-2 14

specific antibody detection in human serum based on magnetic immuno-detection is described and 15

compared to standard ELISA. For magnetic immuno-detection, immunofiltration columns were 16

coated with a SARS-CoV-2 spike protein peptide. SARS-CoV-2 peptide reactive antibodies, spiked 17

at different concentrations into PBS and human serum, were rinsed through immunofiltration 18

columns. Specific antibodies were retained within the IFC and labelled with an isotype specific 19

biotinylated antibody. Streptavidin-functionalized magnetic nanoparticles were applied to label the 20

secondary antibodies. Enriched magnetic nanoparticles were then detected by means of frequency 21

magnetic mixing detection technology, using a portable magnetic read-out device. Measuring 22

signals corresponded to the amount of SARS-CoV-2 specific antibodies in the sample. Our 23

preliminary magnetic immuno-detection setup resulted in a higher sensitivity and broader 24

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

2

detection range and was four times faster than ELISA. Further optimizations could reduce assay 25

times to that of a typical lateral flow assay, enabling a fast and easy approach, well suited for point-26

of-care measurements without expensive lab equipment. 27

Keywords: COVID-19; frequency mixing technology; immunofiltration; magnetic beads, point-of-28

care diagnostic 29

1. Introduction 30

The new Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2), is causing ongoing 31

worldwide infections, leading to an unprecedented pandemic. According to World Health 32

Organization (WHO), it is estimated that up to 82% of people with coronavirus disease 19 (COVID-33

19) are not aware that they are/were infected due to no or very mild symptoms [1]. COVID-19 34

symptoms can be noted comparable to common cold cough, rhinitis or fever up to harsh symptoms, 35

especially at elderly, such as respiratory problems with lung failure and death [2, 3]. Patients with 36

few or no symptoms in particular pose the greatest risk, as they can infect many more people via a 37

droplet infection [4]. Identifying people who were infected and have obtained immunity against 38

SARS-CoV-2 is becoming highly important. Besides valuable epidemiological information regarding 39

distribution and spreading behavior, detection methods would help to manage the currently imposed 40

restrictions and non-pharmaceutical measures. All people with confirmed immunity most probably 41

would have no risk to infect themselves and thus would not represent a risk of infection for others. 42

At a later point, serological assays would also be required to prove and monitor effectivity of 43

vaccination and longevity of the obtained immunity. Fast, cheap and easily applicable on-site testing 44

solutions will thus become increasingly important, but currently only few rapid test systems are 45

available. Lateral-flow-detection (LFD) approaches are easy to handle and results are gained after 10-46

15 min. However they are not quantitative, and their reliability, specificity and sensitivity is much 47

worse than that of lab-based assay formats based on enzyme-linked immunosorbent assay (ELISA). 48

In particular, specificity is a major challenge at currently available serological antibody tests. This 49

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

3

depends to a large extent on the antigen used in the test assays. Enveloped positive-stranded RNA 50

SARS-CoV-2 coronaviruses consist of five structural proteins, the spike glycoprotein (S), envelope 51

protein (E), membrane protein (M), the nucleocapsid protein (N) and a hemagglutinin esterase (HE). 52

The S-protein, a complex folded glycoprotein comprising two regions, S1 and S2, exhibits the highest 53

immunogenicity, has the most important role in host interaction, especially cell entry, and is also the 54

main target for neutralizing antibodies [5]. The proteins M, E and HE are only weakly immunogenic 55

and less suitable as targets for antibody diagnosis. Although the N protein is immunodominant, it is 56

not suitable for the specific analysis of the immune response against SARS-CoV-2 viruses due to its 57

high cross-reactivity with antibodies targeting related coronavirius strains [6, 7]. The company 58

Euroimmun AG, Lübeck, Germany offers two ELISA kits using a genetically modified N-protein 59

variant, which enables a more specific detection of antibodies already ten days after infection. 60

However, for highly specific detection of immune response against SARS-CoV-2, typically the S1 61

subunit of S-protein should be used. Currently, only few vendors offer these specific ELISA formats 62

using the S1 subunit of S-protein for specifically detecting SARS-CoV-2 antibodies [8, 9]. 63

Nevertheless, specific, sensitive and quantitative rapid tests applicable for a decentralized point-of-64

care (PoC) analysis are currently not available. 65

Magnetic Immuno-Detection (MInD) could be a powerful tool for PoC assay performance. MInD 66

employs immunofiltration columns (IFCs) containing a porous polyethylene matrix coated with 67

antigens retaining reactive antibodies from applied samples flushed through IFC by gravity flow. 68

Afterward, secondary antibodies specifically binding to the previously enriched antibodies are 69

applied, and subsequently specially functionalized magnetic particles (MNPs) are added, labelling 70

the secondary antibodies. After a washing step to elute unbound MNPs, IFSc are simply inserted into 71

the detection head of a portable magnetic read-out device in which retained MNPs are then detected 72

by means of frequency magnetic mixing detection technology (FMMD), using a low- and a high-73

frequency magnetic excitation field [10-17]. An alternately oscillating positive and negative magnetic 74

saturation of MNPs with a frequency of 2·f2 of 122 Hz is obtained when exposing the particle to a low 75

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

4

frequency magnetic field of frequency f2 = 61 Hz with an amplitude of a few millitesla [10]. Afterward, 76

the magnetization state of MNPs is probed by the high frequency magnetic excitation field with 77

f1 = 49 kHz. An iron oxide dose-dependent signal is obtained when the resulting mixing frequency 78

signal of f1 + 2·f2 is detected by a Faraday coil. The innermost coil is composed of two adjacent sections, 79

a so-called detection coil directly encircling the IFC, and a reversely wound empty reference coil, so 80

that the directly induced signal from the excitation coils is cancelled while retaining the nonlinear 81

response signal from the MNPs. A detailed description is found in Ref. [10]. An easy application of 82

this technique is guaranteed due to a direct visualization of the resulting measuring signal at the 83

touchscreen of our portable magnetic read-out device. 84

We here present a preliminary MInD proof-of-concept study in comparison to a standard 85

laboratory-based ELISA, demonstrating an improved detection of SARS-CoV-2 specific antibodies 86

spiked in human serum samples. We therefore employed a peptide originating from the SARS-CoV-2 87

Spike protein for IFC coating and antibody enrichment. Our approach might facilitate further 88

optimization to obtain a timely PoC setup for the detection of SARS-CoV-2 specific antibodies in human 89

blood samples. 90

91

2. Materials and Methods 92

2.1 Material and chemicals 93

NaCl, KCl, Na2HPO4 x 12 H20, KH2PO4, Na2(CO3), NaHCO3, Tris-HCL, MgCl2 x 6 H2O and 94

Albumin Fraction V (biotin-free) were acquired from Roth, Karlsruhe, Germany. Peroxidase substrate 95

were purchased from Merck KGaA, Darmstadt, Germany. 96

Coupling buffer was prepared by dissolving 15 mM Na2CO3 and 35 mM NaHCO3 in MilliQ-97

water, and pH was set to 9.6 with glacial acetic acid. Phosphate buffered saline (PBS) was prepared 98

by dissolving 137 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4 x 12 H2O and 1.5 mM KH2PO4 in MilliQ-99

water and setting pH to 7.4 with hydrochloric acid. As washing buffer, PBS with 0.05% (v/v) 100

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

5

Tween-20 was used. ELISA and MInD blocking buffer (BP) was prepared by adding 1% (w/v) 101

albumin fraction V (biotin free) in PBS. Alkaline Phosphatase (AP) buffer was prepared by dissolving 102

100 mM Tris-HCL, 100 mM NaCl and 5 mM MgCl2 x 6 H2O. Buffer was adjusted to pH 9.6 with HCl. 103

Serum sample from normal healthy human was collected in 2016 and stored at -20°C up to usage. 104

Immunofiltration columns (IFC) (ABICAP HP columns) were purchased from Senova 105

Gesellschaft für Biowissenschaft und Technik mbH, Weimar, Germany. High-binding 96-well 106

microtiter plates (article number 655061) were purchased form Greiner Bio-One GmbH, 107

Frickenhausen, Germany. 108

20-aminoacids SARS-CoV-2 spike protein peptide (article number ABIN1382273) derived from 109

the intracellular portion of S2 region of S-protein and corresponding polyclonal rabbit anti-SARS-110

CoV-2 spike protein peptide specific antibody (article number ABIN1030641) were acquired from 111

antibodies-online GmbH, Aachen, Germany. Biotin-SP (long spacer) AffiniPure Goat Anti-Rabbit IgG 112

(biotinylated GaR secondary antibody), Fc fragment specific (article number 111-065-008) as well as 113

Streptavidin-alkaline phosphatase (streptavidin-AP) (article number 016-050-084) were purchased 114

from Jackson ImmunoResearch Europe Ltd. UK. Streptavidin-functionalized magnetic particles with 115

a hydrodynamic diameter of 70 nm (synomag®-D, article number 104-19-701) were purchased from 116

micromod Partikeltechnologie GmbH, Rostock, Germany. 117

118

2.2 ELISA Procedure 119

For ELISA-based SARS-CoV-2 specific antibody detection, all following incubation steps were 120

performed at room temperature for 60 minutes unless stated otherwise. For coating, SARS-CoV-2 121

spike protein peptide was diluted in coupling buffer to a concentration of 2 µg·ml–1 and plated with 122

100 µl per well onto a high-binding 96-well microtiter plate and incubated. After 3 subsequent 123

washing steps with 200 µl PBS-T per well, each well was blocked with BP and incubated. After 124

washing, 100 µl of SARS-CoV-2 spike protein peptide specific antibody in concentrations ranging 125

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

6

from 1.22 ng·ml–1 to 5000 ng·ml–1 diluted in PBS or in human serum acquired in 2016, respectively, 126

were applied onto the microtiter plate. For blank measurements, PBS or human serum without spiked 127

antibody was employed. After incubation, the plate was washed again. Afterward, 100 µl of 128

biotinylated GaR secondary antibody, diluted 1:20,000 in PBS, was added to each well and incubated. 129

After washing three times with PBS-T, 100 µl of streptavidin-AP (diluted 1:1000 in PBS) was added 130

and incubated for 30 minutes. After washing, absorbance was measured using a Tecan Infinite M200 131

microplate reader at 405 nm after application of AP-substrate in AP-buffer and 5 minutes of 132

incubation in dark. 133

134

2.3 Preparation of MInD Immunifiltration Columns 135

For MInD, IFCs were equilibrated as described by Rettcher et al. (2015) [16]. After equilibration, 136

IFCs were coated with 500 µl of 2 µg·ml–1 SARS-CoV-2 spike protein peptide, diluted in coupling 137

buffer. After the solution flushed through by gravity flow, a 60 minutes incubation step was 138

performed. Afterward, IFCs were washed by applying 750 µl PBS. Subsequently, remaining binding 139

sites within the polyethylene filter matrix were blocked by applying twice 750 µl BP and an 140

incubation of 20 minutes after each application. After washing of the columns with 750 µl of PBS, 141

IFCs were ready to use for MInD SARS-CoV-2 spike protein peptide specific antibody detection. Pre-142

coated, ready-to use IFCs could be stored at 4°C for at least several days until they were used for the 143

assay. 144

145

2.4 MInD SARS-CoV-2 Spike Protein Peptide Specific Antibody Detection 146

For proof-of-concept MInD-based detection of antibodies against the SARS-CoV-2 Spike protein 147

peptide, dilutions of an antibody with known specificity for this Spike-protein peptide ranging from 148

1.22 ng·ml–1 to 5000 ng·ml–1 in PBS or in human serum acquired in 2016, respectively, were prepared 149

and applied onto the ready-to use IFCs. While the sample was flushed through the IFC by gravity 150

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

7

flow, SARS-CoV-2 specific antibodies were enriched within the matrix by specific binding to the 151

coated antigen. After 12 minutes of incubation, IFCs were washed with 750 µl PBS and 500 µl of 152

biotinylated GaR (diluted 1:2500 in PBS) were applied onto the columns, binding to retained 153

antibodies. After further incubation of 12 minutes, IFCs were washed again and 500 µl of 60 µg·ml–1 154

70 nm superparamagnetic streptavidin-functionalized magnetic particles were applied and flushed 155

through the IFCs and incubated for 12 minutes. After a final washing with 750 µl PBS, read-out was 156

done by inserting the columns into the portable magnetic reader, detecting the measuring signal in 157

mV as previously described by Rettcher et al. (2015) [16]. 158

159

Data Analysis 160

For ELISA as well as for MInD, data were analyzed using GraphPad Prism 8.0.0, and fitting with 161

a Hill slope was performed. Equations (1) and (2) were used to determine the limit of detection (LOD) 162

or maximum of detection (MOD), respectively, on the signal scale and Equation (3) was used for 163

calculation of those parameters on the concentration scale. SD denotes the standard deviation. 164

𝐒𝐢𝐠𝐧𝐚𝐥𝐋𝐢𝐦𝐢𝐭 𝐨𝐟 𝐃𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧 = 𝐀𝐯𝐞𝐫𝐚𝐠𝐞 𝐬𝐢𝐠𝐧𝐚𝐥𝐁𝐥𝐚𝐧𝐤 + 𝟑 ∙ 𝐒𝐃𝐁𝐥𝐚𝐧𝐤 (1) 165

166

𝐒𝐢𝐠𝐧𝐚𝐥𝐌𝐚𝐱𝐢𝐦𝐮𝐦 𝐨𝐟 𝐃𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧 = 𝐀𝐯𝐞𝐫𝐚𝐠𝐞 𝐬𝐢𝐠𝐧𝐚𝐥𝐬𝐚𝐭𝐮𝐫𝐚𝐭𝐞𝐝 𝐬𝐚𝐦𝐩𝐥𝐞𝐬 − 𝟑 ∙ 𝐒𝐃𝐬𝐚𝐭𝐮𝐫𝐚𝐭𝐞𝐝 𝐬𝐚𝐦𝐩𝐥𝐞𝐬 (2) 167

168

𝐂𝐨𝐧𝐜𝐞𝐧𝐭𝐫𝐚𝐭𝐢𝐨𝐧𝐒𝐚𝐦𝐩𝐥𝐞 = EC50 ∙ (𝐇𝐢𝐠𝐡𝐞𝐬𝐭 𝐒𝐢𝐠𝐧𝐚𝐥−𝐒𝐢𝐠𝐧𝐚𝐥𝐒𝐚𝐦𝐩𝐥𝐞

𝐒𝐢𝐠𝐧𝐚𝐥𝐒𝐚𝐦𝐩𝐥𝐞)

1

𝐻𝑖𝑙𝑙 𝑆𝑙𝑜𝑝𝑒 (3) 169

170

3. Results and Discussion 171

3.1. ELISA-based Calibration Experiments of SARS-CoV-2 Specific Antibody Detection 172

As reference method to our PoC MInD approach, a typical laboratory-based ELISA was performed 173

(Fig 1). After coating of ELISA microtiter plate with SARS-CoV-2 spike protein peptide and blocking 174

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

8

with BSA, SARS-CoV-2 spike protein peptide specific antibody was diluted in the range from 175

1.22 ng·ml–1 to 5000 ng·ml–1 in PBS-buffer or serum and applied into wells. After addition of 176

biotinylated GaR and subsequent labelling with streptavidin-AP, the ELISA plate was read out at 177

405 nm and obtained measuring values were used to generate calibration curves for SARS-CoV-2 178

specific antibody concentrations in PBS (Fig 1, black curve) and in human serum samples (Fig 1, red 179

curve). Blank values determined in PBS and serum were 0.085 AU ± 0.005 AU and 180

0.083 AU ± 0.001 AU, respectively, and were subtracted from sample values. Both curves recorded 181

with antigen diluted in PBS and human blood serum were almost identical. Based on a Hill fit, 182

calibration measurements in PBS with a LOD 3.4 ng·ml–1 and or LOD of 4.0 ng·ml–1 in human serum 183

were obtained (Fig 1). Both calibration measurements show saturation of measuring signals at 184

concentrations with 625 ng·ml–1 up to 5000 ng·ml–1. For calculation of maximum of detection, 185

the average measuring signal of samples with concentrations ranging from 625 ng·ml–1 up to 186

5000 ng·ml–1 was calculated and threefold SD was subtracted. Using equation 3, a MOD of 187

477 ng·ml–1 in PBS and of 312 ng·ml–1 in serum, respectively, could be determined. Following these 188

calibration measurements, it can be concluded, that with our laboratory-based ELISA, SARS-CoV-2 189

specific antibodies can be detected in range of 3.4 ng·ml–1 up to 477 ng·ml–1 in PBS-buffer or from 190

4.0 ng·ml–1 up to 312 ng·ml–1 in human serum samples, respectively. Typical IgG antibody 191

concentrations in human serum are approximately 10 mg·ml–1, and are significantly increasing after 192

antigenic stimulation of immune system [18]. For the whole assay, excluding coating and blocking 193

but including application of sample, secondary antibody, streptavidin-AP and substrate incubation, 194

a procedure time of 161 minutes was needed. This time is comparable to that of commercially 195

available ELISA test kits (e.g. Euroimmun 140 min). Comparing the sensitivity of commercially 196

available ELISA test kits with our standard ELISA will be possible when samples of COVID-19 197

patients are evaluated. A highly sensitive assay could detect antibodies at an early stage of infection, 198

whereas commercially test kits can detect IgG antibodies in patient samples against S1 subunit of 199

SARS CoV-2 S-protein in 75% of samples 10 to 20 days after infection (Euroimmun). 200

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

9

201

Fig 1. ELISA-based detection of SARS-CoV-2 spike protein specific antibody in PBS-buffer (black 202

curve) or spiked in human serum (red curve). Antibody was diluted to concentrations ranging from 203

1.22 ng·ml–1 up to 5000 ng·ml–1 in each matrix and applied onto 2 ng·ml–1 SARS-CoV-2 spike protein 204

peptide coated and BSA blocked microtiter plates. After addition of biotinylated secondary antibody, 205

streptavidin-AP was applied. Limit of detection (LOD) was calculated using non-linear Hill fit 206

(R2=0.997 for PBS-buffer and 0.996 in serum). Assay time of ELISA was 161 minutes. Each data point 207

represents mean ± SD (n = 4 for PBS-buffer and n = 3 for serum). 208

209

3.1. MInD-based Calibration Experiments for SARS-CoV-2 Specific Antibody Detection 210

Same calibration measurements employing dilutions of SARS-CoV-2 specific antibody were 211

done with our PoC MInD-based setup (Fig 2 and 3). Comparable to laboratory-based ELISA, the same 212

dilutions of SARS-CoV-2 spike protein peptide specific antibody in PBS-buffer (Fig 3, black curve) or 213

human serum (Fig 3, red curve) were applied after coating and blocking of IFC with SARS-CoV-2 214

spike protein peptide and BSA. After application of antibody dilutions, a 5-times shorter incubation 215

time of just 12 minutes compared to ELISA was performed. The reduction of incubation time could 216

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

10

be achieved due to a more efficient antibody enrichment. The surface for target binding within the 217

ABICAP immunofiltration column matrix is approximately 40-fold larger compared to the surface of 218

an ELISA well. Furthermore 500 µl of sample was applied and flushed through IFC, compared to just 219

100 µl of sample added to an ELISA microtiter plate well. Afterward, biotinylated GaR was applied 220

onto columns and magnetically labelled with 70 nm streptavidin-functionalized magnetic particles. 221

Again, incubation time of only 12 minutes each could be used due to the increased binding surface. 222

Finally, the IFCs were inserted into our portable magnetic read-out device which can be operated 223

using either an external power adapter or a portable battery, allowing a PoC diagnostic assay 224

procedure. A schematic overview of assay procedure is shown in Fig 2. 225

226

Fig 2. Proof-of-concept MInD assay setup using IFC coated with SARS-CoV-2 antigen. Assay steps 227

and assay time are indicated. IFCs were coated with commercial SARS-CoV-2 S-protein peptide and 228

blocked with BSA. Corresponding antibody was diluted either in PBS or spiked in human serum and 229

applied to IFCs. Biotinylated secondary antibodies were added, followed by application of 230

streptavidin-functionalized MNP. Finally, IFCs were inserted into the portable magnetic read-out 231

device. Measuring signal can be correlated to the amount of antibody in the sample and antibody titer 232

can be determined. Assay time of this preliminary MInD setup was 42 min which is approximately 233

four times faster than ELISA (161 min). 234

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

11

As shown in Fig 3, also with MInD-based approach SARS-CoV-2 specific antibody detection 235

calibration curves could be recorded using PBS-buffer or human serum, respectively. Particularly the 236

comparable calibration measurements in PBS or human serum demonstrate the negligible matrix 237

effect of human serum when using MInD. Compared to standard ELISA, a saturation of measurement 238

signal was observed only at samples with higher concentrations of 2500 ng·ml–1 and 5000 ng·ml–1. 239

Based on this, the average of these two samples was used for calculation of the MOD (Equation 2 and 240

3). By analyzing the range of detection, it can be seen that SARS-CoV-2 spike protein peptide specific 241

antibody can be detected in range of 2.95 ng·ml–1 up to 2040 ng·ml–1 in PBS and from 3.36 ng·ml–1 up 242

to 1810 ng·ml–1, demonstrating an at least 5-fold broader range of quantification in both PBS and 243

human serum with higher sensitivities compared to ELISA and perfectly matching nonlinear fit 244

(R2 =0.997 in PBS and R2 =0.993 in human serum). This significantly increased dynamic range 245

compared to ELISA demonstrates one major advantage enabling improved quantitative 246

measurements. 247

248

Fig 3. Proof-of-concept MInD-based SARS-CoV-2 specific antibody detection. IFCs were coated 249

with commercial 2 µg·ml–1 SARS-CoV-2 spike protein peptide and blocked with BSA. A 250

corresponding antibody was diluted either in PBS (black curves) or spiked in human serum (red 251

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

12

curves) and applied to IFCs. Biotinylated secondary antibodies were added, followed by application 252

of streptavidin-functionalized magnetic particles. Assay time of this preliminary MInD setup was 253

42 minutes (without column preparation). Limit of detection (LOD) was determined by help of 254

non-linear hill fit. Each data point represents mean ± SD (n = 2). 255

256

Furthermore, an approximately 4-fold reduction of assay time with PoC MInD approach 257

compared to ELISA demonstrates the high potential for fast assay procedure. In total, 42 minutes 258

procedure time was needed, resulting in a broader detection range in combination with lower 259

detection limits in both PBS and human serum. A further reduction of assay time could be achieved by 260

checking the required amount of washing steps and analyzing the achieved sensitivity when reducing 261

the incubation time to 5 minutes or less, as described by Rettcher et al. (2015) [16]. Additionally, Rettcher 262

and colleagues demonstrated a reduction of assay time by pre-functionalizing the MNPs with antigen-263

specific antibodies [16]. Here, MNPs can be pre-functionalized with secondary antibody, which would 264

result, in combination with previously mentioned optimization steps, in less assay steps with a 265

comparable assay time as PoC lateral flow assays (less than 20 minutes). 266

In this proof-of-concept experiment, a commercially available SARS-CoV-2 spike protein 267

peptide with corresponding antibody was used. If using this peptide for testing of patient samples, 268

there might be a high risk for false negative assay results, since patients may not have developed 269

antibodies against this peptide. Most preferably, other antigen variants derived from the highly 270

immunogenic S1-subunit of S-protein or a mixture of antigens of SARS-CoV-2 proteins should be 271

used for highly effective and specific enrichment of SARS-CoV-2 targeting antibodies. Especially for 272

demonstration of MInD specificity, control antigens derived from common cold human 273

coronaviruses (hCoV) as hCoV 229E, hCoV NL63, hCoV OC43, hCoV HKU1 or SARS-CoV and 274

MERS-CoV should be tested, confirming the enrichment of only SARS-CoV-2 specific antibodies. 275

Additionally, a multiplex detection of different MNPs, as demonstrated by Achtsnicht et al. (2019), 276

could be implemented for detection of multiple antibody subclasses in one assay [19]. By coupling 277

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

13

different secondary antibodies to MNPs, each type will label a specific antibody isotype, as e.g. IgA 278

and IgG, a course of infection could be visualized as well as the analysis of seroconversion would be 279

enabled. A further optimization of the magnetic read-out device towards a medical-diagnostic device 280

would then fulfill all requirements for use in medical field. To ensure correct sample allocation, the 281

magnetic reader can be equipped with a barcode scanner, and IFCs could be labelled with a patient-282

specific barcode. In combination with multiplex detection of several antibody isotypes, the MInD 283

approach would thus be ideally suited for the use in doctors' surgeries since our MInD assay has less 284

than 10% of the cost for typical ELISA equipment. Especially due to the possibility of performing the 285

PoC assay procedure with a single pipette, and the portable magnetic read-out device which can be 286

battery-operated, the MInD assay can also be used by service providers in the medical field and for 287

testing in elderly peoples’ and nursing homes or at airports, quickly identifying persons with existing 288

immunity. As soon as vaccination against SARS-CoV-2 is available, our approach could be employed 289

to monitor vaccination success and longevity of immunity by determining antibody titers. 290

4. Conclusions 291

We demonstrated for the first time a proof-of-concept MInD-based approach for rapid and highly 292

sensitive SARS-CoV-2 S-protein peptide specific antibody detection in spiked human serum. MInD 293

calibration experiments with a five-fold higher range of detection in combination with higher sensitivity 294

and a four-fold shorter assay time in comparison to a standard ELISA demonstrate the high potential 295

of MInD-based PoC SARS-CoV-2 specific antibody detection in serological samples. By using 296

appropriate SARS-CoV-2 antigens and a multiplex approach for simultaneous detection of e.g. IgA and 297

IgG antibodies reactive against SARS-CoV-2, the state of infection as well as a seroconversion could be 298

analyzed. Especially due to assay speed, low-cost and portable equipment, we conclude that the MInD-299

based assay would be ideally suited for PoC testing, identifying persons with existing immunity. 300

Additionally, our MInD approach could be employed for subsequent analysis of efficiency and for 301

monitoring the longevity of vaccination by determining antibody titers. 302

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

14

Author Contributions: Conceptualization, F.S. and J.P.; methodology, J.P. and N.V.; validation, J.P.; formal 303

analysis, J.P.; Resources, F.S., H.S. and H.J.K; investigation, F.S. and J.P.; data curation, J.P.; writing—original 304

draft preparation, J.P..; writing—review and editing, F.S., H.J.K, H.S. and N.V.; visualization, J.P.; supervision, 305

F.S.; project administration, F.S. 306

Funding: The author received no specific funding for this work. 307

Acknowledgments: The authors would like to thank Max Schubert for his helpful advices and support given in 308

discussions. 309

Competing interests: The authors declare no competing interests. 310

311

References 312

1. Kolifarhood G, Aghaali M, Mozafar Saadati H, Taherpour N, Rahimi S, Izadi N, et al. Epidemiological and 313

Clinical Aspects of COVID-19; a Narrative Review. Arch Acad Emerg Med. 2020;8(1):e41. 314

2. Zhu J, Ji P, Pang J, Zhong Z, Li H, He C, et al. Clinical characteristics of 3,062 COVID-19 patients: a meta-315

analysis. J Med Virol. 2020. 316

3. Wang D, Hu B, Hu C, Zhu F, Liu X, Zhang J, et al. Clinical Characteristics of 138 Hospitalized Patients With 317

2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China. JAMA. 2020. 318

4. Nikolich-Zugich J, Knox KS, Rios CT, Natt B, Bhattacharya D, Fain MJ. SARS-CoV-2 and COVID-19 in older 319

adults: what we may expect regarding pathogenesis, immune responses, and outcomes. Geroscience. 320

2020;42(2):505-14. 321

5. Walls AC, Park YJ, Tortorici MA, Wall A, McGuire AT, Veesler D. Structure, Function, and Antigenicity of 322

the SARS-CoV-2 Spike Glycoprotein. Cell. 2020;181(2):281-92 e6. 323

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

15

6. Agnihothram S, Gopal R, Yount BL, Jr., Donaldson EF, Menachery VD, Graham RL, et al. Evaluation of 324

serologic and antigenic relationships between middle eastern respiratory syndrome coronavirus and other 325

coronaviruses to develop vaccine platforms for the rapid response to emerging coronaviruses. J Infect Dis. 326

2014;209(7):995-1006. 327

7. Khan S, Nakajima R, Jain A, de Assis RR, Jasinskas A, Obiero JM, et al. Analysis of Serologic Cross-328

Reactivity Between Common Human Coronaviruses and SARS-CoV-2 Using Coronavirus Antigen Microarray. 329

bioRxiv. 2020:2020.03.24.006544. 330

8. Zhao J, Yuan Q, Wang H, Liu W, Liao X, Su Y, et al. Antibody responses to SARS-CoV-2 in patients of novel 331

coronavirus disease 2019. Clin Infect Dis. 2020. 332

9. Krüttgen A, Cornelissen CG, Dreher M, Hornef M, Imöhl M, Kleines M. Comparison of four new 333

commercial serologic assays for determination of SARS-CoV-2 IgG. J Clin Virol. 2020. 334

10. Krause H, Wolters, N., Zhanga, Y., Offenhäusser, A., Miethe, P., Meyer, M.H.F. Hartmann, M., Keusgen, 335

M. Magnetic particle detection by frequency mixing for immunoassay applications. J Magn Magn Mater. 336

2007;311(1):436-44. 337

11. Achtsnicht S, Todter J, Niehues J, Teloken M, Offenhausser A, Krause HJ, et al. 3D Printed Modular 338

Immunofiltration Columns for Frequency Mixing-Based Multiplex Magnetic Immunodetection. Sensors (Basel). 339

2019;19(1). 340

12. Achtsnicht S, Neuendorf C, Fassbender T, Nolke G, Offenhausser A, Krause HJ, et al. Sensitive and rapid 341

detection of cholera toxin subunit B using magnetic frequency mixing detection. PLoS One. 2019;14(7):e0219356. 342

13. Meyer MH, Hartmann M, Krause HJ, Blankenstein G, Mueller-Chorus B, Oster J, et al. CRP determination 343

based on a novel magnetic biosensor. Biosens Bioelectron. 2007;22(6):973-9. 344

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint

16

14. Martin H.F. Meyer H-JK, Markus Hartmann, Peter Miethe, Jürgen Oster, Michael Keusgen, . Francisella 345

tularensis detection using magnetic labels and a magnetic biosensor based on frequency mixing. J Magn Magn 346

Mater. 2007;311(1):259-63. 347

15. Meyer MH, Stehr M, Bhuju S, Krause HJ, Hartmann M, Miethe P, et al. Magnetic biosensor for the detection 348

of Yersinia pestis. J Microbiol Methods. 2007;68(2):218-24. 349

16. Rettcher S, Jungk F, Kuhn C, Krause HJ, Nolke G, Commandeur U, et al. Simple and portable magnetic 350

immunoassay for rapid detection and sensitive quantification of plant viruses. Appl Environ Microbiol. 351

2015;81(9):3039-48. 352

17. Pietschmann J, Spiegel H, Krause HJ, Schillberg S, Schroper F. Sensitive Aflatoxin B1 Detection Using 353

Nanoparticle-Based Competitive Magnetic Immunodetection. Toxins (Basel). 2020;12(5). 354

18. Ayres JA, Barraviera B, Calvi SA, Carvalho NR, Peraçoli MTS. Antibody and cytokine serum levels in 355

patients subjected to anti-rabies prophylaxis with serum-vaccination. Journal of Venomous Animals and Toxins 356

including Tropical Diseases. 2006;12:435-55. 357

19. Achtsnicht S, Pourshahidi AM, Offenhausser A, Krause HJ. Multiplex Detection of Different Magnetic 358

Beads Using Frequency Scanning in Magnetic Frequency Mixing Technique. Sensors (Basel). 2019;19(11). 359

360

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 3, 2020. ; https://doi.org/10.1101/2020.06.02.131102doi: bioRxiv preprint