Effect of a Neuropilin-1-Derived Virus Receptor Trap on Enterovirus A71 Infection In Vitro

We discovered that neuropilin 1 (NRP1) is a new receptor candidate to mediate enterovirus A71 (EVA71) into cells. In the engineered form as a decoy receptor, NRP1 was able to recognize and neutralize EVA71 but not enterovirus D68 or coxsackievirus B3 (CVB3). NRP1 recognizes EVA71 through a novel domain on the VP3 capsid protein. The principle in the design, engineering, and refinement of the NRP1-based decoy receptor described in this study represents a general and well-suited antiviral strategy. ABSTRACT We discovered that neuropilin 1 (NRP1) is a new receptor candidate to mediate enterovirus A71 (EVA71) into cells. In the engineered form as a decoy receptor, NRP1 was able to recognize and neutralize EVA71 but not enterovirus D68 or coxsackievirus B3 (CVB3). NRP1 recognizes EVA71 through a novel domain on the VP3 capsid protein. The principle in the design, engineering, and refinement of the NRP1-based decoy receptor described in this study represents a general and well-suited antiviral strategy.

type 1 entry (15,16). Recent studies identified that NRP1, highly expressed in endothelial cells, may facilitate severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) cell entry (17,18). To determine whether NRP1 expression can enhance permissiveness to EVA71, we transfected Chinese hamster ovary (CHO) cells with a plasmid encoding full-length human NRP1 cDNA followed by infection of the CHO-NRP1 cells with recombinant EVA71 encoding a green fluorescence protein (GFP) (EVA71-GFP). We found that the CHO-NRP1 cells stably expressing human NRP1 were permissive to EVA71 infection, whereas EVA71 permissiveness in CHO control cells were almost undetectable (Fig. 1B). Additionally, through a similar strategy, we observed that only NRP1-expressing or scavenger receptor B2 (SCARB2)-expressing, but not nonmodified control BHK21 cells, could be more permissive with EVA71-GFP viruses (Fig. 1C). Because there were more GFP-positive cells in SCARB2-expressing BHK21, SCARB2 was determined to play a more important role for EVA71. To examine the effect of NRP1 expression for EVA71 viral growth kinetics, we infected BHK21, BHK21-SCARB2, and BHK21-NRP1 cells with EVA71 TW/4643/98 (genotype C2) at a multiplicity of infection of 0.01. The viral loads from one run replication cycle (16 h postinfection) of EVA71infected BHK21, BHK21-NRP1, and BHK21-SCARB2 cells were 620 Ϯ 35, 3,267 Ϯ 2,532, and 36,667 Ϯ 2,082 PFU/ml, respectively (Fig. 1D). Significantly, BHK21-NRP1 cells could support EV71 infection and replication as well as BHK21-SCARB2 cells, and the yields of progeny viruses from these modified cells were Ͼ5-fold greater than those from nontransfected cells. Besides, viral proteins in EVA71-infected BHK21-NRP1 or BHK21-SCARB2 were more abundant in Western blot analysis (Fig. 1E). In summary, NRP1 confers susceptibility to EVA71 in BHK21 and CHO cells.
Soluble decoy receptors have become an attractive approach to develop biopharmaceuticals for treatment of autoimmune diseases and abnormal angiogenesis. Prominent examples are soluble tumor necrosis factor-alpha (TNF-␣) receptors, such as etanercept, or VEGF trap, such as aflibercept (19). To improve the pharmaceutical properties of ectodomains of membrane proteins, fragment crystallizable (Fc), derived from immunoglobulin, is often employed as a fusion partner to produce chimeric decoy receptor fusion proteins with a prolonged serum half-life in vivo through the Fc receptor-mediated recycling mechanism (20). During the preparation of this work, recombinant angiotensin-converting enzyme 2 (ACE2), as a virus receptor trap, was shown to be effective in controlling SARS-CoV-2 infection and viral growth (21). Currently, human recombinant soluble ACE2 is being tested as a drug candidate to treat SARS-CoV-2 infection in a phase 2 trial (ClinicalTrials registration no. NCT04335136). In this study, we designed and produced NRP1-ECD-(a1b2)-hFc fusion proteins (where hFc is human IgG fragment crystallizable) ( Fig. 2A). Recombinant NRP1-ECD-(a1b2)-hFc fusion protein was successfully expressed and purified, and its molecular weight and purity was revealed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 2B).
This proof-of-concept study demonstrated that NRP1 can mediate EVA71 binding and infection in cells. The novel decoy receptor derived from NRP1 can neutralize EVA71 infectivity in vitro. Recently, the b1b2 domain of NRP1 was found to play a role in facilitating SARS-CoV-2 cell entry (17,18). Our findings highlight that pharmaceutical optimization of NRP1-ECD-(b1b2)-hFc, a virus receptor trap, would be a valid strategy for developing broad-spectrum antiviral agents to combat emerging virus infection.
Transduction efficiencies in CHO cells expressing NRP1. The CHO-S cell line was purchased from Thermo Fisher Scientific, Inc. A commercial expression vector-encoding hNRP1 (Sino Biological) was used to establish CHO-NRP1 cells. Stable NRP1-CHO cells were selected after hygromycin B treatment. EVA71 (strain C2 4643/TW/98) with GFP (EVA71-GFP), which expressed GFP on infection, was kindly supplied by Hung-Yao Ho from the Department of Medical Biotechnology and Laboratory Science, Chang Gung University. The reporter virus was generated by engineering the GFP gene between the reducing [DTT(ϩ)] conditions. (C) An indirect ELISA was performed by coating wells with 20 g/ml formalin-inactivated EVA71, and the bound fusion proteins were detected using HRP-conjugated goat anti-hFc antibody. Both NRP1-ECD-(a1b2)-hFc and NRP1-ECD-(b1b2)-hFc bound to EVA71 with similarly high affinities (10 to 11 nM). All data were calculated from the averaged duplicate values and analyzed using GraphPad Prism. (D) An ELISA-based competitive binding assay was performed by coating wells with 10 g/ml formalin-inactivated EVA71. Then, 0.2 M recombinant scavenger receptor B2 (SCARB2) fusion protein was competitively bound with or without NRP1-ECD-hFc (0.04 to 1 M), and binding was detected by using an HRP assay. (E) An EVA71 VP3 peptide library containing 47 peptides was used in an overlapping-peptide ELISA for epitope mapping, and bound antibodies were detected using an HRP assay. 5= untranslated region and VP4 gene of the EVA71 genome. We detached the CHO-NRP1 or CHO cells in phosphate-buffered saline (PBS) and mixed them with EVA71-GFP (1 ϫ 10 6 PFU/ml) for 1 h at 37°C. We washed the cells twice with PBS, suspended them in CDM4PerMab medium plus 2% fetal bovine serum (FBS), and incubated them at 37°C. After 72 h, we observed the GFP-positive cells under a microscope. Transduction efficiencies in BHK21 cells expressing NRP1 or SCARB2. The BHK21 cell line was purchased from Bioresource Collection and Research Center (BCRC, Taiwan). A commercial expression vector-encoding NRP1 with an N-Myc tag (Sino Biological) was used to establish BHK21-NRP1 cells, and a commercial expression vector-encoding hSCARB2 with qa C-OFPSpark tag (Sino Biological) was used to establish BHK21-SCARB2 cells. GFP-positive cells were observed under a microscope 72 h after EVA71-GFP infection.
Production of recombinant fusion proteins. All recombinant proteins with Fc domains of human IgG1 (GenBank accession no. AEV43323.1) were produced from CHO cells and purified by using a protein A-derived affinity medium (GE Healthcare). A synthetic CHO codon-optimized target-protein cDNA sequence was cloned into a mammalian cell expression vector and then transfected into CHO cells (Thermo Fisher Scientific). Culture of CHO cells was performed using serum-free CDM4PerMab medium (GE Healthcare). The pSecTag2/Hygro mammalian expression vector (Thermo Fisher Scientific) containing the hinge and hFc was used to express NRP1-ECD-hFc fusion proteins. The NRP1-ECD-(a1b2) region encodes both the a1a2 (amino acids [aa] 22 to 272) and b1b2 (aa 273 to 586) domains, whereas NRP1-ECD-(b1b2) contains only the b1b2 domain (aa 273 to 586) (NCBI reference sequence NP_001019799.1).
ELISA analysis. A recombinant viral receptor competitive binding assay, based on ELISA, with NRP1-ECD-hFc fusion proteins was performed by coating Nunc MaxiSorp ELISA plates with 10 g/ml formalin-inactivated EVA71 (EVA71 TW/73/12). Subsequently, 0.2 M recombinant SCARB2 fusion protein with tags (His tag and human Fc; Sino Biological) was prepared with or without different concentrations of NRP1-ECD-(a1b2)-hFc or NRP1-ECD-(b1b2)-hFc. Bound recombinant SCARB2 fusion protein was detected through incubation with the HRP-conjugated mouse anti-His tag antibody (BioLegend). The percentage difference in competitive inhibition was calculated using the formula 100 Ϫ ([B/A] ϫ 100), where A represents the optical density at 450 nm (OD 450 ) of the recombinant SCARB2 fusion protein control group without NRP1-ECD-hFc competitors, and B represents the OD 450 of the experimental group with NRP1-ECD-hFc competitors.
Overlapping-peptide ELISA-based epitope mapping was performed by coating Nunc MaxiSorp ELISA plates with 20 g/ml synthetic peptides. A library of EVA71 (TW/2086/ 98) VP3 region overlapping synthetic peptides was kindly provided by Yen-Hung Chow and Chia-Chyi Liu (The National Health Research Institutes) (25). Forty-seven overlapping peptides were used for epitope mapping, and each peptide was 15 amino acids long, with 10-amino acid overlaps between sequential peptides. Bound antibodies were detected through incubation with the HRP-conjugated goat anti-human IgG-Fc fragment cross-adsorbed antibody (Bethyl Laboratories).
Virus preparation and cell-based neutralization assay. EVA71 TW/73/12 (genotype C4), EVD68 TW/02795/14, and CVB3 were isolated from the Chang Gung Memorial Hospital clinical virology laboratory. In brief, 96-well plates were seeded with 3 ϫ 10 4 RD cells/well (ATCC CCL-136) in Dulbecco's modified Eagle medium with 10% FBS and incubated overnight at 37°C. The NRP1-ECD-hFc fusion proteins were incubated with a 100ϫ 50% tissue culture infective dose (TCID 50 ) of EVA71 for 1 h at 37°C. After adsorption, the infected cells were covered with a medium containing 2% FBS. The infected cells were then incubated at 37°C for 64 h. The plates were fixed with formaldehyde and then stained with crystal violet. Well density at 570 nm was measured. Each experiment was performed in triplicate and repeated at least twice. The 50% effective concentration (EC 50 ) was calculated using the formula (Y Ϫ B)/(A Ϫ B) ϫ (H Ϫ L) ϩ L, where Y is half the mean OD at 570 nm (OD 570 ) of the control cells without the protein, B is the mean OD 570 of wells with the protein dilution nearest to and ϽY, A is the mean OD 570 of wells with the protein dilution nearest to and ϾY, and L and H are the protein concentrations for B and A, respectively.
Virus titration and viral protein detection in infected cells. The EVA71 TW/ 4643/98 (genotype C2) used in the virus titration study was kindly provided by Jen-Ren Wang, National Cheng Kung University, Taiwan (26). Virus titers were determined by plaque assay using RD cells after a single replication cycle (16 h postinfection) in BHK21, BHK21-NRP1, and BHK21-SCARB2 cells. Intracellular viral proteins were detected by Western blotting using the anti-EVA71 monoclonal antibody MAB979 (Merck KgaA).