A novel in vitro system of supported planar endosomal membranes (SPEMs) reveals an enhancing role for cathepsin B in the final stage of Ebola virus fusion and entry

Abstract

Ebola virus (EBOV) causes a hemorrhagic fever with fatality rates up to 90%. The EBOV entry process is complex and incompletely understood. Following attachment to host cells, EBOV is trafficked to late endosomes/lysosomes where its glycoprotein (GP) is processed to a 19-kDa form, which binds to the EBOV intracellular receptor Niemann-Pick type C1. We previously showed that the cathepsin protease inhibitor, E-64d, blocks infection by pseudovirus particles bearing 19-kDa GP, suggesting that further cathepsin action is needed to trigger fusion. This, however, has not been demonstrated directly. Since 19-kDa Ebola GP fusion occurs in late endosomes, we devised a system in which enriched late endosomes are used to prepare supported planar endosomal membranes (SPEMs), and fusion of fluorescent (pseudo)virus particles is monitored by total internal reflection fluorescence microscopy. We validated the system by demonstrating the pH dependencies of influenza virus hemagglutinin (HA)-mediated and Lassa virus (LASV) GP-mediated fusion. Using SPEMs, we showed that fusion mediated by 19-kDa Ebola GP is dependent on low pH, enhanced by Ca²⁺, and augmented by the addition of cathepsins. Subsequently, we found that E-64d inhibits full fusion, but not lipid mixing, mediated by 19-kDa GP, which we corroborated with the reversible cathepsin inhibitor VBY-825. Hence, we provide both gain- and loss-of-function evidence that further cathepsin action enhances the fusion activity of 19-kDa Ebola GP. In addition to providing new insights into how Ebola GP mediates fusion, the approach we developed employing SPEMs can now be broadly used for studies of virus and toxin entry through endosomes. IMPORTANCE Ebola virus is the causative agent of Ebola virus disease, which is severe and frequently lethal. EBOV gains entry into cells via late endosomes/lysosomes. The events immediately preceding fusion of the viral and endosomal membranes are incompletely understood. In this study, we report a novel in vitro system for studying virus fusion with endosomal membranes. We validated the system by demonstrating the low pH dependencies of influenza and Lassa virus fusion. Moreover, we show that further cathepsin B action enhances the fusion activity of the primed Ebola virus glycoprotein. Finally, this model endosomal membrane system should be useful in studying the mechanisms of bilayer breaching by other enveloped viruses, by non-enveloped viruses, and by acid-activated bacterial toxins.

Keywords: Ebola virus; Lassa; acid activated toxins; endosomal receptors; endosomes; enveloped virus; filovirus; genome entry; hemorrhagic fever viruses; influenza; non-enveloped virus; viral membrane fusion.

FIG 1

Preparation of SPEMs. (A) Immunoblots of gradient fractions probed for different organelle markers: early endosomes (EEA1 and Rab5), late endosomes, endolysosomes (NPC1 and Lamp1), plasma membrane (Na/K ATPase), Golgi apparatus (GS28), mitochondria (SDHA), and endoplasmic reticulum (ER) (calnexin). These are representative immunoblots from two independent endosome enrichment preparations. Preparation 1 was analyzed for EEA1, Rab5, NPC1, Lamp1, Na/K ATPase, and GS28. Preparation 2 was analyzed for calnexin and SDHA. For the latter, the NP/PNS and FR1/FR2 samples were separated by intervening lanes; hence, in the immunoblots of calnexin and SDHA, the NP/PNS and FR1/FR2 lanes were spliced, as indicated by the black vertical line. FR1, 7%–14% optiprep interface; FR2, 14%–25% optiprep interface. (B) Cartoon showing preparation of SPEMs. A lipid monolayer was transferred onto a clean quartz slide by immersing and pulling it through a monolayer of a lipid mix of brain PC, cholesterol, and DPS in the ratio 77:20:3. Next, the lipid monolayer covered slide was assembled into a watertight flow cell, and endolysosomes from FR1 were injected into the flow cell. The endolysosomes spontaneously spread on the supported monolayer to form SPEMs. (C) FRAP experiment on SPEMs. Carboxy-fluorescein-phosphatidylethanolamine (PE) was included in the monolayer prior to preparation of the bilayer. Total fluorescence intensities during patterned FRAP experiments from SPEMs as well as from lipid-only bilayers of brain PC and cholesterol (4:1). At least 10 regions on four independently prepared lipid-only bilayers and SPEMs were sampled to determine the average values reported. Fluorescence is scaled to the intensity range before and immediately after bleaching. DPS, 1,2-dimyristoyl-sn-gycero-3-phosphatidylethanolamine-PEG3400-triethoxysilane; FR1, fraction 1; FR2, fraction 2; FRAP, fluorescence recovery after photobleaching; NP, nuclear pellet; PC, phosphatidylcholine; PNS, post-nuclear supernatant; SDHA, succinate dehydrogenase complex subunit A.

FIG 2

Influenza virus (FPV) fusion to SPEMs. (A) Cartoon depicting DiD-labeled FPV fusion to SPEMs. FPV was labeled with DiD during viral production; hence, the membrane label (cyan) is present in both leaflets. (B) Intensity traces of particles undergoing docking only (23 traces from pH 7.4 condition), hemifusion (7 intensity traces from pH 5.5 condition), and full fusion (21 intensity traces from pH 5.5 condition) were aligned and averaged (red traces). The shaded area represents standard deviation. (C) pH dependence of hemifusion (black) and full fusion (pink) of DiD-labeled FPV to SPEMs within a 3-min time period. Movies were recorded at a frame rate of 100 ms for a minimum of 2,000 frames. The total number of docked particles was quantitated for each condition. Each data point represents events observed on one separately prepared SPEM. Error bars indicate standard error. Statistical comparison was performed using Welch’s two-tailed t test. **P < 0.01. All comparisons not shown are not significant.

FIG 3

Lassa virus GP pseudovirus fusion to SPEMs. (A)Twenty intensity traces of fusion of HIV pseudovirions bearing Lassa GP, labeled with Atto488-DMPE in the outer membrane leaflet, were averaged. (Particles in this average curve were taken from the pH 5.5, no Lamp1 condition.) (B) Immunoblots of gradient fractions from HEK293T WT and Lamp1 KO cells probed for different organelle markers as denoted in the legend of Fig. 1. (C) Fraction of particles bearing Lassa GP that underwent lipid mixing with SPEMs with or without Lamp1. The virus particles were introduced into the flow cell at the indicated pH. Recording began immediately after input of viral particles into the flow cell. Data from five to nine SPEMs were averaged under each condition. Data from individual SPEMs are shown in Fig. S5C and D. Error bars indicate standard error. Welch’s two-tailed t test is shown above data. *P < 0.05. All comparisons not shown are not significant.

Fig 4

Ebola virus GP pseudovirus fusion to SPEMs. (A) Intensity traces of DiD-labeled EBOV GPcl pseudovirions undergoing hemifusion (23 intensity traces from pH 5.5, EDTA condition) and full fusion (21 intensity traces from pH 5.5, EDTA condition). Ebola pseudoviruses were labeled with 1-µM DiD during viral production so as to label both membrane leaflets. Traces were aligned and averaged (red traces). The shaded area represents standard deviation. (B) DiD-labeled HIV particles pseudotyped with Ebola GPΔ, treated with thermolysin to generate 19-kDa GP1 (GPcl) or untreated (GPΔ), fusing with SPEMs at pH 7.4 and pH 5.5 in the presence of 1-mM EDTA or pH 5.5 in the presence of Ca²⁺ or Mg²⁺ (0.5-mM CaCl₂ or MgCl₂). Each data point represents the average percent fusion of approximately 70–200 docked particles observed on each SPEM at 2.5 min post-adding low pH buffer. Statistical comparison was performed using Welch’s two-tailed t test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. All comparisons not shown are not statistically significant. (C through E) Cumulative distribution function for full fusion and hemifusion of EBOV GPcl pseudovirus with SPEMs in the presence or absence of added cathepsin B (0.4 µg/mL) (C) or cathepsin L (0.4 µg/mL) (D). Recordings began when cathepsin B in low pH buffer was pumped into the flow cell chamber (0 min) as well as at 5, 10, 15, 20, 25, and 30 min after. Residual Ca²⁺ was present from thermolysin treatment done to obtain cleaved GP. Error bars indicate standard error. (E) Unpaired two-tailed t test for the 30-min data point in the cumulative distribution functions is shown: *P < 0.05, ****P < 0.0001. All comparisons not shown are not statistically significant. (F through H) Cumulative distribution function for full fusion (F) and hemifusion (G) of EBOV GP_cl pseudovirus particles with SPEMs in the presence or absence of 100-µM E-64d. Recordings began when low pH buffer with or without E-64d was injected into the flow cell chamber (0 min) as well as 5, 10, 15, 20, 25, and 30 min after. Data points in panels C, D, F and G are averages obtained from four to six SPEMs under each condition. Error bars indicate standard error. (H) Unpaired two-tailed t test for the 30-min data point in the cumulative distribution functions is shown: **P < 0.01, ***P < 0.001. All comparisons not shown are not statistically significant.

Fig 5

Effect of a reversible cathepsin inhibitor, VBY-825, on Ebola virus GP pseudovirus fusion to SPEMs. (A through C) Cumulative distribution functions for full fusion (A) and hemifusion (B) of EBOV GP_cl pseudovirus particles with SPEMs in the presence or absence of 1-µM VBY-825. Recording began when low pH buffer, with or without VBY-825, was injected into the flow cell chamber (0 min) as well as at 5, 10, 15, 20, 25, and 30 min after. To test for reversibility of the inhibition, 1-µM VBY-825/low pH was introduced into the flow cells, and movies were collected at times 0, 5, 10, and 15 min, after which VBY-825 was washed out with low pH buffer (red arrow), and additional movies were collected at the 20-, 25-, and 30-min timepoints. All buffers contained 0.1% dimethyl sulfoxide (DMSO). Data points in panels are averages obtained from three to six SPEMs under each condition. Error bars indicate standard error. (C) Unpaired two-tailed t test for the 30-min data point in the cumulative distribution functions is shown. *P < 0.05, **P < 0.01, ***P < 0.001. All comparisons not shown are not statistically significant.