Two key residues in ephrinB3 are critical for its use as an alternative receptor for Nipah virus

Abstract

EphrinB2 was recently discovered as a functional receptor for Nipah virus (NiV), a lethal emerging paramyxovirus. Ephrins constitute a class of homologous ligands for the Eph class of receptor tyrosine kinases and exhibit overlapping expression patterns. Thus, we examined whether other ephrins might serve as alternative receptors for NiV. Here, we show that of all known ephrins (ephrinA1-A5 and ephrinB1-B3), only the soluble Fc-fusion proteins of ephrinB3, in addition to ephrinB2, bound to soluble NiV attachment protein G (NiV-G). Soluble NiV-G bound to cell surface ephrinB3 and B2 with subnanomolar affinities (Kd = 0.58 nM and 0.06 nM for ephrinB3 and B2, respectively). Surface plasmon resonance analysis indicated that the relatively lower affinity of NiV-G for ephrinB3 was largely due to a faster off-rate (K(off) = 1.94 x 10(-3) s(-1) versus 1.06 x 10(-4) s(-1) for ephrinB3 and B2, respectively). EphrinB3 was sufficient to allow for viral entry of both pseudotype and live NiV. Soluble ephrinB2 and B3 were able to compete for NiV-envelope-mediated viral entry on both ephrinB2- and B3-expressing cells, suggesting that NiV-G interacts with both ephrinB2 and B3 via an overlapping site. Mutational analysis indicated that the Leu-Trp residues in the solvent exposed G-H loop of ephrinB2 and B3 were critical determinants of NiV binding and entry. Indeed, replacement of the Tyr-Met residues in the homologous positions in ephrinB1 with Leu-Trp conferred NiV receptor activity to ephrinB1. Thus, ephrinB3 is a bona fide alternate receptor for NiV entry, and two residues in the G-H loop of the ephrin B-class ligands are critical determinants of NiV receptor activity.

Figure 1. Soluble NiV-G Binds to EphrinB3 with Lower Affinity than EphrinB2

(A) 1.0 μg/ml, 0.1 μg/ml, and 0.01 μg/ml of the indicated ephrin-Fc fusion proteins were allowed to bind to soluble NiV-G-coated plates in an ELISA format (see Materials and Methods). The amount of ligand bound was detected colorimetrically using an antihuman Fc antibody conjugated to horseradish peroxidase. One representative experiment out of three is shown. Data are averages of triplicates ± standard error (SE). (B) EphrinB2 and B3 stably transfected CHO-pgsA745 cells (CHO-B2 and CHO-B3, respectively) were used to measure NiV-G-Fc cell surface binding. Increasing concentrations of NiV-G-Fc were added to either CHO-B2 cells (dashed line with squares) or CHO-B3 cells (solid line with triangles), and binding was assessed by flow cytometry using R-phycoerythrin-conjugated anti-Fc antibodies. Regression curves were generated as described in Materials and Methods. Each data point is an average ± SE from three experiments. (C) Surface plasmon resonance (BIAcore 3000) measured the binding kinetics of NiV-G-Fc to both ephrinB2-Fc and ephrinB3-Fc in response units (RU). NiV-G-Fc was immobilized to a CM5 sensor chip via an amide coupling procedure, and increasing concentrations of ephrinB2-Fc and ephrinB3-Fc were flowed as analyte over the sensor chip. One representative experiment out of two is shown. (D) K _d, K _on (association-rate), and K _off (dissociation-rate) were determined by fitting the binding chromatogram data from (C) with BIAcore evaluation software (version 3.1) using the 1:1 Langmuir binding model.

Figure 2. Pseudotyped and Live NiV Use EphrinB2 and B3 for Cellular Entry

(A) Ephrin expression was measured by flow cytometry on CHO-pgsA745 parental cells (CHO) and CHO-pgsA745 cells stably expressing ephrinB1, B2, and B3 (CHO-B1, CHO-B2, and CHO-B3). To bind the CHO cell lines, 10 μg/ml of EphA2, 10 μg/ml EphB3-Fc, and 1 nM of NiV-G-Fc were used, and the amount of binding was detected by flow cytometry as in Figure 1B. Data are representative of three experiments. (B) NiV-F and G glycoproteins were pseudotyped onto a VSV-ΔG-Luc core virus (NiV-VSV-ΔG-Luc) and used to infect parental CHO-pgsA745 (CHO), CHO-B1, CHO-B2, and CHO-B3 cells. Entry of the indicated dilutions of NiV-VSV-ΔG-Luc viruses was measured by quantifying Renilla Luc activity according to manufacturer's directions. Relative light units (RLU) were acquired and quantified on a Veritas luminometer. Data are shown as averages of triplicates ± standard deviation of a representative experiment. In three independent experiments, viral entry into CHO-B3 cells was reduced by 21%, 28%, and 46%, respectively, compared to CHO-B2 cells (p = 0.05, paired t-test). (C) The listed MOIs of live NiV were used to infect the indicated cell lines (10⁵ cells per infection). Foci of syncytia were observed 24 h postinfection. Vero E6 cells are fully permissive for NiV infection and were used as positive control cells. Note the larger number of syncytia seen on CHO-B2 versus CHO-B3 cells.

Figure 3. EphrinB2 and B3 Bind NiV-G at an Overlapping Site

NiV-VSV-ΔG-Luc pseudotyped viruses were used to infect CHO-B2 and CHO-B3 cells in the presence of the indicated amounts of ephrinB1, B2, and B3-Fc fusion proteins (B1-Fc, B2-Fc, and B3-Fc, respectively). Entry was measured as in Figure 2A. Data are the average of triplicates ± standard deviation, and one representative experiment of three is shown.

Figure 4. The Leu–Trp Residues Present in the G–H Loop of EphrinB2 and B3 Are the Critical Determinants of NiV-G Binding

(A) Sequence alignment of human (hu), mouse (ms), and rat (rt) ephrinB1, B2, and B3 ectodomains using the Jotun Hein algorithm (DNAstar Megalign software). Six residues in the ephrin B-class ectodomain reveal solvent-exposed amino acids [22] that contain identical residues in both ephrinB2 and B3 but different residues in ephrinB1 (open box). Examination of the ephrin binding loop (G–H loop) indicates the L–W residues in ephrin B2 and B3 are replaced by Y–M residues in ephrinB1 (filled box). (B) Ephrin-Fc mutants were created by substituting the L–W residues present in ephrinB2 and B3 with Y–M residues using site-directed mutagenesis (B2YM-Fc and B3YM-Fc). Conversely, the Y–M residues in ephrinB1 were exchanged for the L–W residues (B1LW-Fc); 10 nM, 1 nM, and 0.1 nM of both wild-type (B1, B2, and B3) and mutant (B1LW, B2YM, and B3YM) ephrin-Fc proteins were tested for their ability to bind NiV-G-HA in an ELISA. The amount of binding was measured the same as in Figure 1A. The data are averages of three experiments done in triplicates ± standard error.

Figure 5. The Leu–Trp Residues in G–H loop of EphrinB3 Are Necessary for Pseudotyped NiV Entry

(A) The percentage of ephrin cell surface expression (CSE) was measured by flow cytometry on CHO-pgs745 parental cells (CHO) and CHO-pgs745 cells stably expressing both full-length wild-type ephrins (B1, B2, and B3) and mutant ephrins (B1LW, B2YM, and B3YM); 10 μg/ml of EphB3-Fc (solid bar) and 1 nM of NiV-G-Fc (open bar) were used to bind the CHO cell lines, and the amount of binding was detected the same as in Figure 1B. The data are an average of triplicates ± standard deviation (SD). (B) The same CHO cell lines used above were seeded at 10⁵ cells per well and infected with pseudotyped NiV-VSV-ΔG-Luc virus. The amount of entry was detected as in Figure 2A. One representative experiment of three is shown, and data are an average of triplicates ± SD. In three independent experiments, the viral entry into B2YM cells was reduced by 45%, 68%, and 85%, respectively, compared to wild-type B2 cells (p < 0.03, paired t-test).