key: cord-284102-rovyvv45
authors: Wagner, Teresa R.; Kaiser, Philipp D.; Gramlich, Marius; Becker, Matthias; Traenkle, Bjoern; Junker, Daniel; Haering, Julia; Dulovic, Alex; Schweizer, Helen; Nueske, Stefan; Scholz, Armin; Zeck, Anne; Schenke-Layland, Katja; Nelde, Annika; Strengert, Monika; Walz, Juliane S.; Ruetalo, Natalia; Schindler, Michael; Schneiderhan-Marra, Nicole; Rothbauer, Ulrich
title: NeutrobodyPlex - Nanobodies to monitor a SARS-CoV-2 neutralizing immune response
date: 2020-09-28
journal: bioRxiv
DOI: 10.1101/2020.09.22.308338
sha: 
doc_id: 284102
cord_uid: rovyvv45

As the COVID-19 pandemic escalates, the need for effective vaccination programs, diagnosis tools and therapeutic intervention ever increases. Neutralizing binding molecules have become important tools for acute treatment of COVID-19 and also provide a unique possibility to monitor the emergence and presence of a neutralizing immune response in infected or vaccinated individuals. Here we identified 11 unique nanobodies (Nbs) with high binding affinities to the SARS-CoV-2 spike receptor domain (RBD). Of these, 8 effectively block the RBD:ACE2 interface. Via competitive binding analysis and detailed epitope mapping, we grouped all Nbs into 3 sets and demonstrated their neutralizing effect. Combinations from different sets showed a profound synergistic effect by simultaneously targeting different epitopes within the RBD. Finally, we established a competitive multiplex binding assay (“NeutrobodyPlex”) enabling the detection of neutralizing antibodies in serum of infected patients. Overall, our Nbs have high potential for prophylactic and therapeutic options and provide a novel approach to screen for a neutralizing immune response in infected or vaccinated individuals, helping to monitor immune status or guide vaccine design.

7 phycoerythrin (PE)-labeled streptavidin after stringent washing. Additionally, a non-specific Nb 135 (GFP-Nb, negative control) and two inhibiting mouse antibodies (positive controls) were 136 analyzed 21 . Data obtained by this multiplex binding assay showed that 8 of the 10 analyzed 137

Nbs inhibit ACE2 binding to isolated RBD, S1 domain and homotrimeric spike. IC50 values 138 calculated for inhibition of ACE2:RBD interaction ranges between 0.5 nM for NM1228 and 38 139 nM for NM1229 (Figure 3) . Notably, IC50 values obtained for the most potent inhibitory Nbs 140 NM1228 (0.5 nM), NM1226 (0.85 nM) and NM1230 (2.12 nM) are highly comparable to IC50 141 values measured for the mouse IgGs (MM43: 0.38 nM; MM57: 3.22 nM). Additionally, the 142 assay revealed that all Nbs except NM1224, show a similarly strong inhibitory effect of ACE2 143 binding to all tested antigens. NM1224 seems to exclusively inhibit RBD:ACE2 interaction and 144

does not prevent binding of ACE2 to either the homotrimeric Spike or the S1 domain. 145 146

After identifying RBD-specific Nbs which have an inhibitory effect on ACE2 binding, we 148 investigated the relative location of their epitopes within the RBD. Firstly, we first performed 149 epitope binning experiments of Nb combinations using biolayer interferometry. After coating 150 sensors with biotinylated RBD, a Nb was loaded until binding saturation was reached, followed 151 by a short dissociation step to remove excess Nb. A second Nb from a different family was 152 then exposed to the RBD-Nb-complex. Using this approach, we identified Nbs which recognize 153 overlapping and non-overlapping epitopes on RBD (Figure 4, Supplementary Figure 2) . As 154 expected Nbs with only minor differences in their CDR3 (NM1221, NM1222 and NM1230, Nb-155

Set 2) were suggested to recognize an identical or highly similar epitope as they cannot bind 156 simultaneously to RBD. Our analysis revealed that Nbs with highly diverse CDR3s such as 157 NM1228, NM1226, NM1227 and NM1229 could not bind simultaneously, suggesting that these 158

Nbs recognize similar or at least overlapping epitopes. As a result, we clustered these diverse 159

Nbs in Nb-Set 1. Overall, we identified five distinct Nbs-Sets, comprising at least one candidate 160 targeting a different epitope within the RBD compared to any member of a different Nb-Set 161 (Figure 4) . 162

8

Next, we performed Hydrogen-Deuterium Exchange Mass Spectrometry (HDX-MS) with the 164 most potent inhibitory Nbs selected from the different Nb-Sets. This allowed us to more 165 precisely locate their binding sites at the surface of RBD and compare with the RBD:ACE2 166

interface. Both members of Nb-Set1, NM1226 and NM1228, interacted with the RBD at the 167 back/ lower right site (Back View, Figure 5) . Notably, the binding site of NM1226 does not 168 encompass amino acid residues involved in the RBD:ACE2 interface. In contrast, NM1228 169

(Nb-Set1) as well as NM1230 (Nb-Set2) contacted the RBD at amino acid residues overlapping 170 with the RBD:ACE2 binding interface, whereas NM1230 additionally covers parts of the spike-171 like loop region on one edge of the ACE2 interface at the top front/ lower left side (Front View, 172 which did not contact any amino acid residues involved in the RBD:ACE2 interface but rather 179 binds to the opposite site (Front View, Figure 5 ). Comparing the data from epitope binning 180 with the HDX-MS results, provides structural insights into the mechanism by which non-181 competing pairs of Nbs can simultaneously bind the RBD. Interestingly, the combination of 182 NM1228 (Nb-Set1) with NM1230 (Nb-Set2) shows near complete coverage of the ACE2 183 interface (Figure 5) whereas the observed inhibitory effect of NM1226 might be due to steric 184 hindrance. From these findings, we proposed that the combination of Nb-Set1 with Nb-Set2 185 might act synergistically on the inhibition of the interaction between RBD and ACE2. 186

After identification of Nbs which inhibit the RBD:ACE2 interaction biochemically, we employed 189 a cell-based viral infection assay to test for their neutralization potency. To this end, human 190 9 Caco-2 cells were co-incubated with the icSARS-CoV-2-mNG strain and serial dilutions of the 191 inhibitory Nbs NM1224, NM1226, NM1228 and NM1230. 48 h post-infection neutralization 192 potency was determined via automated fluorescence-microscopy of fixed and nuclear-stained 193 cells (Supplementary Figure 4) . Percentage of the infection rate following Nb treatment 194 normalized to a non-treated control was plotted and IC50 values were determined via sigmoidal 195 inhibition curve fits. Overall, data obtained from the multiplex binding assay and the viral 196 infection assay were broadly consistent. Representatives of Nb-Set1, NM1226 and NM1228, 197 showed the highest neutralization potency with IC50 values of ~15 nM and ~7 nM followed by 198 NM1230 (~37 nM) and NM1224 (~256 nM). As expected, NM1223 (Nb-Set3) was not found to 199

Considering that Nbs targeting diverse epitopes within the RBD:ACE2 interface are beneficial 201 in both reducing viral infectivity and preventing mutational escape, we next combined the most 202 potent inhibitory and neutralizing candidates derived from Nb-Set1 (NM1226, NM1228) and 203

Nb-Set2 (NM1230) and examined their response in both the multiplex binding assay and viral 204 infection assay. In the multiplex binding assay the combination of NM1226 and NM1230 205

showed an increased effect in competing with ACE2 binding to RBD illustrated by a IC50 of 206 0.42 nM which is 2-or 5-fold lower compared to treatment with individual NM1226 or NM1230, 207 respectively (Figure 7 A) . Notably, the IC50 measured for the combination of NM1228 and 208 NM1230 did not exceed the IC50 identified for NM1228 alone indicating that NM1228 by its own 209 has a very high inhibiting effect (Figure 7 A). When we tested both combinations in the viral 210 infection assay, we observed significantly improved effects in both as illustrated by an IC50 of 211 ~4 nM for the combination NM1226 and NM1230 and ~3.5 nM for NM1228 and NM1230 212 (Figure 7 B, Supplementary Figure 5 ). From these findings we conclude, that a combinatorial 213 treatment with two Nbs targeting different epitopes within the RBD:ACE2 interaction site is 214 beneficial for viral neutralization. context, multiple studies have convincingly shown that neutralizing antibodies preferable bind 226 to the RBD domain and sterically inhibit viral entry via ACE2 1,3 . From this, we can assume that 227 our RBD Nbs covering large parts of the RBD:ACE2 interface might be suitable to monitor the 228 emergence and presence of neutralizing antibodies in patients. To test this hypothesis, we set 229 up a high-throughput competitive binding assay, termed NeutrobodyPlex, by combining our 230 most potent neutralizing Nb combinations with a recently developed, automatable multiplex 231 immunoassay (Figure 8 A) 20 . We incubated our previously generated color-coded beads 232 comprising RBD, S1 domain or homotrimeric spike with serum samples from patients or non-233 infected individuals, in addition to dilution series of the combinations NM1226/ NM1230 or 234 NM1228/ NM1230 and used this to detect patient-derived IgGs bound to the respective 235 antigens. Depending on the Nb concentration, neutralizing antibodies targeting the RBD:ACE2 236 interaction site within the serum samples are displaced resulting in a reduction of the 237 detectable signal (Figure 8 A) . 238

When analyzing RBD specific IgGs from serum samples, we detected a distinct signal 239 reduction in the presence of increasing Nb concentrations for all tested samples (Figure 8 To further demonstrate that our approach is able to determine the presence of IgGs targeting 246 the RBD:ACE2 interaction site in detailed resolution, we highlight here the effect of competing 247

IgGs could be observed when measuring binding to RBD, however using the S1 domain as 249 target antigen distinct differences between both serum samples became visible. While #289 250 comprise a substantial fraction of IgGs addressing the RBD:ACE2 interface also presented by 251 the S1 domain, in sample #265 IgGs binding to additional epitopes of the S1 domain cover the 252 detectable signal reduction derived from displaced IgGs (Figure 8 For functional analysis we employed a recently developed in vitro multiplex binding assay 20 to 291 monitor the replacement of ACE2 as the natural ligand from binding to RBD, S1 domain or 292 homotrimeric spike upon addition of RBD-specific Nbs. With this assay, we were able to identify 293 8 inhibiting Nbs targeting those spike-derived antigens. Interestingly, IC50 values obtained for 294 inhibitory Nbs on RBD and homotrimeric spike show a higher correlation compared to IC50 295 values obtained for the S1 domain. Based upon detailed epitope mapping, we grouped our 296

Nbs in 5 different Nb-Sets. 3 of those Nb-Sets, comprise inhibitory Nbs which were shown to 297 target different epitopes within the RBD:ACE2 interaction site. We confirmed the neutralizing 298 potency of those Nbs in a cell-based viral infection assay using fully intact SARS-CoV-2. 299

Through this, we noted that the measurable viral neutralization effect of the individual Nbs 300 strongly correlates to the data obtained from the biochemical screen, which demonstrates that 301 13 the multiplex binding assay as presented is highly relevant and suitable to identify virus 302 neutralizing binders. As a result, we modified our previously described multiplex immunoassay 303 (MULTICOV-AB, 20 ) and developed a novel diagnostic test called NeutrobodyPlex to monitor 304 the presence and the emergence of neutralizing antibodies in serum samples of SARS-CoV-2 305 infected individuals. Using combinations of high affinity Nbs covering the RBD:ACE2 interface, 306

we were able to directly and specifically displace IgGs present in serum samples from these 307 particular RBD epitopes. According to previous studies, human IgGs addressing those 308 epitopes were classified as neutralizing antibodies 1,17,18 . In our NeutrobodyPlex, we further 309 demonstrated that such neutralizing antibodies can be detected best using the RBD. Larger 

Expression constructs For bacterial expression of Nbs, sequences were cloned into the 331 pHEN6 vector 28 , thereby adding a C-terminal 6xHis-tag for IMAC purification as described 332

previously 29,30 . The pCAGGS plasmids encoding the stabilized homotrimeric spike protein and 333 the receptor binding domain (RBD) of SARS-CoV-2 were kindly provided by F. Krammer 19 . 334

The cDNA encoding the S1 domain (aa 1 -681) of the SARS-CoV-2 spike protein was obtained 335 by PCR amplification using the forward primer S1_CoV2-for 5´-CTT CTG GCG TGT GAC 336 CGG -3´ and reverse primer S1_CoV2-rev 5´ -GTT GCG GCC GCT TAG TGG TGG TGG with high-confidence identification (q-value ≤ 0.01) were included to the list. Peptides with 499 overlapping mass, retention time and charge in Nb and antigen digest, were manually 500 removed. The deuterated samples were recorded in MS mode only and the generated peptide 501 list was imported into HDExaminer v2.5.0 (Sierra Analytics, Modesto, CA, USA). Deuterium 502 uptake was calculated using the increase of the centroid mass of the deuterated peptides. 503 HDX could be followed for 79% of the RBD amino acid sequence. The calculated percentage 504 deuterium uptake of each peptide between RBD-Nb and RBD-only were compared. Any 505 peptide with uptake reduction of 5% or greater upon Nb binding was considered as protected. 506 507 Cell culture Caco-2 (Human Colorectal adenocarcinoma) cells were cultured at 37°C with 5% 508 CO2 in DMEM containing 10% FCS, 2 mM l-glutamine, 100 μg/ml penicillin-streptomycin and 509 1% NEAA. Results from bead-based multiplex ACE2 competition assay are shown for the three SARS-582

CoV-2 spike-derived antigens, RBD, S1 and homotrimeric spike. ACE2 bound to the respective 583 antigen was detected. For each Nb, a dilution series from 2.106 µM to 0.123 nM is shown in 584 the presence of 80 ng/mL ACE2. MFI signals were normalized to the maximal signal per 585 antigen as given by the ACE2-only control. IC50 values were calculated from a four-parametric 586 sigmoidal model and are displayed for each Nb and antigen. Data is presented as mean +/-587 SD of three technical replicates (n =3). 588 

Isolation of potent SARS-CoV-2 neutralizing antibodies and 667 protection from disease in a small animal model

Structural basis for the recognition of SARS-CoV-2 by full-length human

Human neutralizing antibodies elicited by SARS-CoV-2 infection

Characterization of the receptor-binding domain (RBD) of 2019 novel 674 coronavirus: implication for development of RBD protein as a viral attachment inhibitor 675 and vaccine

Nanobodies: natural single-domain antibodies

Structural Basis for Potent Neutralization of Betacoronaviruses by 679

Neutralizing nanobodies bind SARS-CoV-2 spike RBD and block 681 interaction with ACE2

An ultra-high affinity synthetic nanobody blocks SARS-CoV-2 infection 683 by locking Spike into an inactive conformation. bioRxiv

Humanized Single Domain Antibodies Neutralize SARS-CoV-2 by 686

Targeting Spike Receptor Binding Domain. bioRxiv

An alpaca nanobody neutralizes SARS-CoV-2 by blocking receptor 689 interaction. bioRxiv

Affinity Nanobodies Block SARS-CoV-2 Spike Receptor Binding Domain Interaction 692 with Human Angiotensin Converting Enzyme. bioRxiv, 2020

Fast isolation of sub-nanomolar affinity alpaca nanobody against the 695

Spike RBD of SARS-CoV-2 by combining bacterial display and a simple single-step 696 density gradient selection. bioRxiv

Multivalent Nanobody Cocktails for Highly Efficient SARS-699

A potent neutralizing nanobody against SARS-CoV-2 with inhaled delivery 702 potential. bioRxiv

Spike mutation pipeline reveals the emergence of a more transmissible 704 form of SARS-CoV-2. bioRxiv

Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G

Increases Infectivity of the COVID-19 Virus

Potent Neutralizing Antibodies against SARS-CoV-2 Identified by High-710

Throughput Single-Cell Sequencing of Convalescent Patients' B Cells

A neutralizing human antibody binds to the N-terminal domain of the Spike 713 protein of SARS-CoV-2

A serological assay to detect SARS-CoV-2 seroconversion in humans

Going beyond clinical routine in SARS-CoV-2 antibody testing -A 717

multiplex corona virus antibody test for the evaluation of cross-reactivity to endemic 718 coronavirus antigens. medRxiv

Quantum Dot-Conjugated SARS-CoV-2 Spike Pseudo-Virions 721

Enable Tracking of Angiotensin Converting Enzyme 2 Binding and Endocytosis

Evaluation of nine commercial SARS-CoV-2 immunoassays

Convergent antibody responses to SARS-CoV-2 in convalescent 726 individuals

A translational multiplex serology approach to profile the prevalence 728 of anti-SARS-CoV-2 antibodies in home-sampled blood. medRxiv

A high-throughput neutralizing antibody assay for COVID-19 731 diagnosis and vaccine evaluation

Speed up to find the right ones: rapid discovery of functional nanobodies

A SARS-CoV-2 surrogate virus neutralization test based on antibody-736 mediated blockage of ACE2-spike protein-protein interaction

Selection and identification of single domain antibody fragments from camel heavy-740 chain antibodies

Modulation of protein properties in living cells using nanobodies

A versatile nanotrap for biochemical and functional studies with 745 fluorescent fusion proteins

Targeting and tracing antigens in live cells with fluorescent 748 nanobodies

SARS-CoV-2 Seroconversion in Humans: A Detailed Protocol for 750

Antigen Production, and Test Setup

Deuterium Exchange Mass Spectrometry to Study Protein Complexes

Optimization of Feasibility Stage for Hydrogen/Deuterium 756

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the 759 ACE2 receptor

S1 domain or homotrimeric spike of SARS-CoV-2 was incubated with Nb combinations 442 (concentrations ranging from 1.26 µM to 0.08 nM for each Nb) and serum samples of 443 convalescent SARS-CoV-2 patients and healthy donors at a 1:400 dilution. As positive control 444 and maximal signal detection per sample, serum only was included and as negative control for 445Nb binding a SARS-CoV-2-unspecific GFP nanobody (1.26 µM) was used. To compare Nb 446 performance, the inhibiting mouse antibody (40591-MM43) was added in concentrations of 447 0.17 µM to 0.08 nM. Bound serum IgGs were detected via anti-human-IgG-PE as previously 448 and fragments mass tolerance were set to 6 ppm and 0.05 Da, respectively. No enzyme 495 selectivity was applied, however, identified peptides were manually evaluated to exclude 496 peptides originated through cleavage after arginine, histidine, lysine, proline and the residue