Fig 1.
Establishing the pseudotyping of LV particles with DARPin-displaying NiV glycoproteins.
(A) Schematic drawing of modified NiV envelope proteins. The glycoprotein G variant consisting of differently truncated cytoplasmic tails (CT), transmembrane domain (TM), ectodomain (ED), the fused EpCAM-specific DARPin Ac1 and a His tag is shown in the top row. Below, the schematic drawing of the fusion protein (F) cytoplasmic tail variants is shown. (B) Unconcentrated screening titers (t.u./ml) of all nine combinations of G and F protein variants titrated on CHO-EpCAM cells (n = 3; mean ± standard deviations (SD) are shown; ****, P<0.0001; ns, not significant by one-way ANOVA with Tukey's multiple comparisons test). (C) Surface expression of EpCAM-targeted G variants on HEK-293T cells transiently transfected with the corresponding expression plasmids (empty curves) compared to mock transfected cells (filled curves) as determined by flow cytometry. Cells were stained with PE-coupled anti-His antibody. One representative out of three experiments is shown. For quantitative data see S1 Fig. (D) Western blot analysis of LV particles for incorporation of GEpCAM, Gc∆33EpCAM, and Gc∆34EpCAM and the three different F variants (full length F, Fc∆22 and Fc∆25). For generation of LVs used for Western blot analysis, in contrast to other vector productions, F protein with C-terminally fused AU1 tag was used in order to allow detection via an anti-AU1 antibody. Incorporation of the G variants was detected via an anti-His antibody. 2.5x1010 particles per sample were applied. NiV-GHis/F (GHis-LV) and Gc∆34His/Fc∆22 (NiVwt-LV) pseudotyped LVs as well as concentrated supernatant of mock transfected cells (mock) and concentrated supernatant of cells transfected with the gag/pol encoding plasmid pCMV∆R8.9 only (bald-LV) served as controls. M indicates the marker lane. For quantitative data see S2 Fig. (E) In order to optimize titers, the ratio of the amounts of the plasmids encoding the Gc∆34EpCAM and the Fc∆22 protein was varied for vector production in HEK-293T cells as indicated. The produced vector stocks were titrated on CHO-EpCAM cells, and their relative titers, normalized to that obtained after transfection of the 1:3 ratio, is shown.
Fig 2.
Mutation of the NiV glycoprotein to ablate natural receptor recognition.
(A) Surface representation of top-view of NiV-G. G is shown as dimer by modeling its monomer crystal structure (Protein Data Bank (PDB) ID: 3D11) on the crystal structure of the Hendra virus G dimer (PDB ID: 2X9M) using PyMOL. The binding site for ephrin-B2 is depicted in blue [26]. Residues mutated in G to screen for their potential to ablate natural receptor tropism are shown in red. (B) Six different single mutations and the combinations E501A+W504A (Gc∆34EpCAMmut2.1), Q530A+E533A (Gc∆34EpCAMmut2.2), or E501A+W504A+Q530A+E533A (Gc∆34EpCAMmut4) were introduced into Gc∆34EpCAM. Unconcentrated vector stocks generated with the mutated G proteins were titrated on CHO-EpCAM (white bars; negative for natural NiV receptors) and U87-MG cells (black bars; positive for NiV receptors). NiV-LVs pseudotyped with Gc∆34/Fc∆22 having the natural NiV tropism (Gc∆34) served as control. Arrows indicate titers <6x102 t.u./ml. Statistics refer to unmutated Gc∆34EpCAM. Titers of EpCAM-targeted LVs on CHO-EpCAM cells were not statistically different. (n = 4; mean ± standard deviations (SD) are shown; *, P<0.1 **, P<0.01 by one-way ANOVA with Dunnett's multiple comparisons test). (C-D) Binding of ephrin-B2 (C) and B3 (D) to NiV-G mutants is shown. HEK-293T cells were transfected either mock or with plasmids encoding the indicated G protein variants and then incubated with 1 μg/ml recombinant Fc-ephrin-B2 or -B3 prior to staining against the Fc-tag using FITC coupled anti-Fc antibody. The binding efficiencies of the different mutants to the receptors are shown as MFI (mean fluorescence intensity) values (n = 3; mean ± standard deviations (SD) are shown). (E) Concentrated vector stocks of Gc∆34EpCAMmut4/Fc∆22 pseudotyped LV vectors were generated and particle size was analyzed via single nanoparticle tracking analysis (NTA). Particle size measurement of one representative out of three independent stocks is shown (black). As control, concentrated vectors stocks of VSV-LV (blue) and NiVwt-LV (red) were analyzed. The mean size ± SD of the main peak out of three measurements of each particle type is indicated. (F) Electron microscopy of concentrated LV particles pseudotyped with Gc∆34EpCAMmut4/Fc∆22 proteins. The white arrowhead points to the NiV glycoproteins on the particle surface. Scale bar: 100 nm.
Fig 3.
Selectivity of NiVmutEpCAM-LV for EpCAM+ cells.
(A) Cell entry is EpCAM dependent but independent of ephrin-B2. Representative flow cytometry plots out of three independent experiments of CHO-K1, CHO-EpCAM, CHO-ephrin-B2 cells, and of a mixed culture composed of CHO-EpCAM and CHO-ephrin-B2 (1:1 ratio) monitored 72 h after transduction with NiVmutEpCAM-LV, NiVwt-LV or VSV-LV (MOI of 1). EpCAM expression was detected by an APC-coupled human EpCAM specific antibody. (B) To ascertain stability of transduction with the EpCAM-targeted vector, CHO-EpCAM cells were cultivated for further 30 days after transduction with the indicated MOIs. The percentage of GFP-positive cells was determined by flow cytometry at the indicated time points. One representative out of three independent experiments is shown.
Fig 4.
Receptor usage of NiVmutEpCAM-LV.
To determine the receptor usage of NiVwt-LV (A) in comparison to NiVmutEpCAM-LV (B) a competition assay was performed by incubating the vector particles (MOI 0.4) for 1 h at 4°C with increasing amounts of the entire extracellular domain of human ephrin-B2 (black lines), ephrin-B3 (blue lines), human EpCAM (red lines) or murine EpCAM (grey lines). Following incubation, cells were transduced and analyzed for GFP expression by flow cytometry 72 h post transduction. Data are normalized to transduction efficiency measured without pre-incubation with recombinant protein (n = 3). (C) CHO-EpCAM and CHO-ephrin-B2 cells (1x104) were pre-treated with increasing amounts of bafilomycin A1 for 30 minutes prior to transduction of cells with NiVmutEpCAM-LV and NiVwt-LV at an MOI of 0.4. VSV-LV served as a control as cell entry of this vector is described to be pH-dependent. Relative transduction rates compared to cells transduced in absence of bafilomycin A1 were determined by flow cytometry 72 h post transduction (n = 3).
Fig 5.
Expanding the system to additional target receptors.
(A) Surface expression of NiV G proteins (blue line) targeted to four different receptors was compared to that of the corresponding MV H protein counterparts (red line). All expression plasmids encoding the different constructs were transfected into HEK-293T cells. Surface expression was analyzed after 48 hours using a His-tag-specific antibody (PE-labeled). Mock transfected cells (filled curves) served as negative control. One representative out of three experiments is shown. For quantitative data see S3 Fig. (B) Western blot analysis of NiVmutEpCAM-LV, NiVmutCD8-LV, NiVmutCD20-LV, and NiVmutHer2-LV. For generation of the vectors used for Western blot analysis, F protein C-terminally tagged with the AU1 immunological tag was used to allow detection via the anti-AU1 antibody. Incorporation of the G variants was detected via an anti-His antibody. 2.5x1010 particles per sample were used. Mock transfected cells (mock) as well as bald particles without glycoproteins (bald-LV) served as controls. In addition, particles pseudotyped with full-length His-tagged G and AU1 tagged F (GHis-LV) as well as particles pseudotyped with Gc∆34His/Fc∆22-AU1 (NiVwt-LV) were used. For quantitative data see S4 Fig. (C) Titers of receptor-targeted NiV-LVs and their MV-LV counterparts. Unconcentrated stocks of EpCAM-targeted vectors were titrated on CHO-EpCAM cells, CD20-targeted vectors on Raji, CD8 targeted-vectors on Molt4.8 and Her2/neu-targeted vectors on SK-OV-3 cells. (n = 4; mean ± standard deviations (SD) are shown; *, P<0.1; ***, P<0.001; ****, P<0.0001 by unpaired t-test). (D) Fold change in titers (t.u./ml) determined by normalizing the titers of NiV glycoprotein based LVs to those of the corresponding MV glycoprotein based LVs. (E) Concentration of vector stocks from (C) by centrifugation. (F) Number of transducing units per 108 physical particles of NiV and MV glycoprotein based LVs. Particle numbers were determined by single nanoparticle tracking analysis (NTA) (n = 4; mean ± standard deviations (SD) are shown; *, P<0.1; ***, P<0.001; ns, not significant by unpaired t-test).
Fig 6.
Neutralization of NiV glycoprotein pseudotyped LVs.
CHO-EpCAM or CHO-ephrin-B2 cells were transduced with NiVmutEpCAM-LV (blue), MVEpCAM-LV (red), NiVwt-LV (black), or VSV-LV (grey) at an MOI of 0.4 after incubation with serial dilutions of pooled human serum (IVIG) for 2 h at 37°C. After 72 h, GFP+ cells were determined by flow cytometry. The number of GFP+ cells relative to the untreated control is shown (n = 3).
Fig 7.
Selective transduction of CD8+ human PBMC.
(A) Freshly isolated human PBMC were activated for three days and then transduced with NiVmutCD8-LV, NiVwt-LV or VSV-LV at an MOI of 2, respectively. Representative flow cytometry plots out of three independent experiments are shown. GFP fluorescence was measured via flow cytometry at day 5, 10 and 17 after transduction. CD8+ cells were stained with APC coupled anti-CD8 antibody. (B) Percentage of GFP positive cells within the CD8+ population was measured at day 5, 10 and 17 after transduction (n = 3).
Fig 8.
Characterization of NiVmutHer2-LV.
(A) Surface expression of Gc∆34mutHer2 and Hc∆18mutHer2 in comparison to Gc∆34His on HEK-293T cells transiently transfected with plasmids encoding the different glycoproteins compared to mock transfected cells as determined by flow cytometry. Cells were stained with PE coupled anti-His antibody. Mean fluorescence intensities of three independent measurements are shown (n = 3; mean ± standard deviations (SD) are shown; ns, not significant by unpaired t-test). See S6 Fig for exemplary raw data. (B) Binding of recombinant Her2/neu to the engineered glycoproteins. Gc∆34Her2mut4, Hc∆18mutHer2 and Gc∆34His were expressed in HEK-293T cells, incubated for 1 h at 4°C with 1 µg/ml recombinant Fc-Her2/neu prior to staining against the Fc-tag using FITC coupled anti-Fc antibody. Mean fluorescence intensities of three experiments are shown (n = 3; mean ± standard deviations (SD); ns, not significant by unpaired t-test). See S6 Fig for exemplary raw data. (C) Western blot analysis of NiV- and MV-based Her2/neu targeted LVs. 2.5x1010 (1), 1.25x1010 (0.5), 6.25x109 (0.25) and 3.125x109 (0.125) particles were applied per sample, respectively. The glycoproteins G and H were detected by anti-His antibody. The NiV F glycoprotein was detected via the AU1-tag specific antibody. One representative out of three Western blots is shown. The central lane between the NiV and MV samples containing the molecular weight marker was removed. (D) Incorporation levels of Gc∆34Her2mut4 and Hc∆18mutHer2. Three independently generated stocks of NiVmutHer2-LV and MVHer2-LV were subjected to Western blot analysis applying four different particle numbers as shown in (C). Average chemiluminescence values for the glycoproteins G and H were then normalized to those of p24 (n = 3 for all NiVmutHer2-LV dilutions and MVHer2-LV dilutions 1 and 0.5; n = 2 for dilutions 0.25 and n = 1 for dilution 0.125 of MVHer2-LV; mean ± standard error of the mean (SEM) are shown). (E) Bafilomycin A1 sensitivity of NiVmutHer2-LV. NiVmutHer2-LV, NiVwt-LV and VSV-LV were titrated on SK-OV-3 cells in presence or absence of 20 nM bafilomycin A1. Relative titers of vectors in presence of bafilomycin A1 to untreated control are shown (n = 3). (F) Exchanging the DARPin 9.29 with alternative Her2/neu-specific targeting domains. Mean fluorescence intensities of surface expression of Gc∆34Her2mut4 variants in which DARPin 9.29 was replaced by DARPins 9.01, 9.16, 9.26, H14R, G3 or the scFv 4D5++ after transient transfection of HEK-293T cells with the corresponding expression plasmids compared to mock transfected cells as determined by flow cytometry. Cells were stained with PE-coupled anti-His antibody (n = 4; mean ± standard deviations (SD) are shown). Examples of representative flow cytometry plots are shown in S7 Fig. (G) Western blot analysis of the different NiV glycoprotein based vectors targeted to Her2/neu. The G variants were detected via an anti-His antibody, and F by AU1-tag specific antibodies. 2.5x1010 particles per sample were applied. Mock transfected cells (mock) as well as bald particles without glycoproteins (bald-LV) served as controls. In addition, particles pseudotyped with full-length His-tagged G and AU1 tagged F (GHis-LV) as well as particles pseudotyped with Gc∆34His/Fc∆22-AU1 (NiVwt-LV) were used. For quantitative data see S8 Fig. (H) Titers of concentrated vector stocks of the different Her2/neu specific NiV-LVs and of MVHer2-LV as determined on SK-OV-3 cells (n = 3; mean ± standard deviations (SD) are shown).
Fig 9.
CD117- and GluA4-targeted NiV-LVs.
(A) Surface expression of CD117 (blue line) and CD117short (red line) on stably expressing HT1080 cells compared to the parental cell line (filled curve) as determined by flow cytometry. Cells were stained with PE-coupled CD117 antibody. One representative out of four experiments is shown. See S10 Fig for quantitative data. (B) Surface expression of GluA4 (blue line) and GluA4short (red line) on stably expressing HT1080 cells compared to the parental cell line (filled curve) as determined by flow cytometry. Cells were stained with PE-coupled myc-tag antibody. One representative out of three experiments is shown. See S10 Fig for quantitative data. (C) Binding of recombinant SCF to CD117 and CD117short. Fc-SCF was produced in HEK-293T cells by transient transfection. HT1080, HT1080-CD117 and HT1080-CD117short were incubated for 1 h at 4°C with the same volumes of recombinant Fc-SCF prior to staining against Fc-tag using FITC coupled anti-Fc antibody. One representative out of three experiments is shown. See S10 Fig for quantitative data. (D) Titers of concentrated stocks of NiVmutCD117-LV and NiVmutGluA4-LV. NiVmutCD117-LV was titrated on HT1080-CD117 and HT1080-CD117short cells (n = 4; mean ± standard deviations (SD) are shown; **, P<0.01by unpaired t-test). NiVmutGluA4-LV was titrated on HT1080-GluA4 and HT1080-GluA4short cells (n = 3; mean ± standard deviations (SD) are shown; ***, P<0.001 by unpaired t-test).
Fig 10.
Position of the binding site on the targeted receptor determines cell entry.
(A) Three-dimensional structures of the targeted receptors and the positions of their binding sites relative to the cell membrane. Surface representation of the extracellular domains of ephrin-B2 (PDB ID: 2VSM), EpCAM (PDB ID: 4MZV), CD20 (PDB ID: 3PP4), CD8 (PDB ID: 1CD8), Her2/neu (PDB ID: 1N8Z), CD117 (PDB ID: 2E9W), CD117short (adapted from PDB ID: 2E9W). For GluA4 the crystal structure of the closely related GluA2 including the transmembrane domain is shown (PDB ID: 3KG2). For GluA4short, the amino terminal domain (ATD) of GluA4 (PDB ID: 4GPA) is shown. Non-crystalized membrane-proximal amino acids (aa) of undefined structure are indicated with blue dots, each dot representing about 20 residues (ephrinB2: 64 aa; CD20: 25 aa (the structure of the small 6 aa loop is not available either); CD8α: 47 aa; Her2/neu: 23 aa; CD117: 17 aa; CD117short: 24 aa, GluA4short: 24 aa). When available, the structure of the complex between the target receptor and the targeting domain (red) is shown: Her2/neu with bound DARPin-9.29 (D-9.29, PDB ID: 4HRL) and DARPin-G3 (D-G3, PDB ID: 4HRN) (adapted from [37]). The CD117 ligand, SCF, is shown bound to CD117 and CD117short. For the natural NiV receptor ephrin-B2 the complex with the bound NiV-G monomer is shown. (B) Molecular model for the distance effect and its implication for NiV-mediated membrane fusion. In absence of receptor binding, F is in its prefusion state with the fusion peptide (light blue) being covered within the globular head (left). Upon attachment of G to its cell surface receptor, conformational changes are induced resulting in the projection of the fusion peptide followed by its insertion into the cell membrane (top right). If the attached binding site on the receptor is too far away from the cell membrane, the fusion peptide cannot insert and cell entry will not proceed (bottom right). Model adapted from [2] and [50].