Human parvovirus B19 interacts with globoside under acidic conditions as an essential step in endocytic trafficking

Abstract

The glycosphingolipid (GSL) globoside (Gb4) is essential for parvovirus B19 (B19V) infection. Historically considered the cellular receptor of B19V, the role of Gb4 and its interaction with B19V are controversial. In this study, we applied artificial viral particles, genetically modified cells, and specific competitors to address the interplay between the virus and the GSL. Our findings demonstrate that Gb4 is not involved in the binding or internalization process of the virus into permissive erythroid cells, a function that corresponds to the VP1u cognate receptor. However, Gb4 is essential at a post-internalization step before the delivery of the single-stranded viral DNA into the nucleus. In susceptible erythroid Gb4 knockout cells, incoming viruses were arrested in the endosomal compartment, showing no cytoplasmic spreading of capsids as observed in Gb4-expressing cells. Hemagglutination and binding assays revealed that pH acts as a switch to modulate the affinity between the virus and the GSL. Capsids interact with Gb4 exclusively under acidic conditions and dissociate at neutral pH. Inducing a specific Gb4-mediated attachment to permissive erythroid cells by acidification of the extracellular environment led to a non-infectious uptake of the virus, indicating that low pH-mediated binding to the GSL initiates active membrane processes resulting in vesicle formation. In summary, this study provides mechanistic insight into the interaction of B19V with Gb4. The strict pH-dependent binding to the ubiquitously expressed GSL prevents the redirection of the virus to nonpermissive tissues while promoting the interaction in acidic intracellular compartments as an essential step in infectious endocytic trafficking.

Fig 1. Gb4 is essential at a post-internalization step and before the delivery of the viral genome into the nucleus for replication.

(A) Detection of VP1u cognate receptor (VP1uR) and Gb4 in UT7/Epo wild-type (WT) and Gb4 knockout (KO) cells by immunofluorescence (IF). Cells were incubated with either recombinant VP1u labelled with anti-FLAG antibody or anti-Gb4 antibody, washed, fixed and stained with secondary antibodies for confocal microscopy. B19V uptake (1h at 37°C) was detected with the antibody 860-55D against intact capsids, while progeny virus (72h at 37°C) was detected with the antibody 3113-81C against viral capsid proteins. (B) B19V uptake (1h) and infection (72h) in cells pre-incubated with recombinant VP1u (VP1u ΔC126) (S1 Fig) for 30 min at 4°C to block the VP1uR. (C) Attachment (1h at 4°C) and uptake (1h at 37°C) of fluorescently labelled (Atto 488) MS2 bacteriophage capsids with conjugated recombinant VP1u ΔC126 from B19V (MS2-Atto⁴⁸⁸-VP1u) (S1 Fig). As a control, unconjugated fluorescently labelled MS2 bacteriophage capsids were used (MS2-Atto⁴⁸⁸). Nuclei were stained with DAPI. (D) NS1 mRNA quantification by RT-qPCR after transfection of genomic ssDNA (1h and 48h) in WT and Gb4 KO cells. The results are presented as the mean ± SD of three independent experiments. ns, not significant. (E) Detection of capsid protein expression (3113-81C; green) after transfection of genomic ssDNA (48h) in WT and Gb4 KO cells.

Fig 2. In the absence of Gb4, internalized B19V is arrested in the endosomal compartment.

(A) Co-localization of incoming B19V capsids (860-55D; green) with endo-lysosomal markers (M6PR and Lamp1; red) in UT7/Epo WT and Gb4 KO cells at 30 min and 3h pi. (B) Quantitative analysis of scattered cytoplasmic foci (B19V capsid signal) not colocalizing with endocytic markers at 3h pi (S3 Fig). Unpaired Students t-test with Welch’s correction (not assuming same SD) was used for statistical comparison. ****, p<0.0001. (C) Co-localization of incoming B19V capsids (860-55D; green) with MS2-Atto⁵⁹⁴-VP1u in UT7/Epo WT and Gb4 KO cells at 3h pi. (D) Co-localization of B19V capsids (860-55D; green) with a cis-Golgi marker (GM130; red) in WT cells at 3h pi.

Fig 3. B19V interacts with soluble and membrane Gb4 in a pH-dependent manner.

(A) Hemagglutination of human RBCs (0.5% in 100 μl PiBS) by B19V or MVM (5x10⁹) at different pH values. (B) Hemagglutination of RBCs with a 20-fold increase of B19V particles at pH 7.4 and 6.3. (C) Quantitative analysis of B19V binding to RBCs at pH 7.4 and 6.3 in the presence or absence of GSLs (Gb3 or Gb4). Virus (5x10⁹) was incubated directly with RBCs (0.5% in 100 μl PiBS) at the indicated pH or pre-blocked with 5x10¹⁴ molecules (Gb3 or Gb4) for 1h at pH 6.3 prior to incubation with RBCs. After 1h at room temperature, the erythrocytes were washed four times with the corresponding buffer and the viral DNA was extracted and quantified. The results are presented as the mean ± SD of three independent experiments. ****, p<0.0001; ns, not significant. (D) Hemagglutination inhibition test in the presence or absence of different amount (0.013–1.6 μM) of GSLs (Gb3 or Gb4) at pH 6.3. (E) Hemagglutination inhibition test in the presence of anti-Gb4 or non-specific IgY antibodies at pH 6.3 (antibodies were incubated with RBCs 1h before adding the virus). (F) Hemagglutination of RBCs (0.5% in 100 μl PiBS) by VLPs (5x10⁹) at different pH values. (G) Detection of VLPs bound to RBCs at different pH values by Western blot using antibody 3113-81C. (H) Hemagglutination inhibition test in the presence or absence of different amount (0.013–1.6 μM) of Gb3 or Gb4 at pH 6.3.

Fig 4. Determination of the optimal conditions required for B19V and Gb4 interaction.

(A) Determination of the optimal pH for B19V binding to Gb4. RBCs (0.5% in 100 μl PiBS) were incubated with B19V (5x10⁹) at pH values ranging from 8.5 to 5.5. After 1h, cells were washed in the corresponding buffer, and viral DNA was extracted and quantified by qPCR. (B) Hemagglutination at neutral (7.4) or acidic (6.3) pH was carried out in the presence of divalent cations (Ca²⁺ or Mg²⁺), chelating agents (5 mM EGTA or EDTA) or in minimal essential medium (MEM). B19V was incubated with the different buffers for 1h before incubation with the erythrocytes. HA, hemagglutination assay.

Fig 5. pH acts as an affinity switch to regulate binding and dissociation between B19V and Gb4.

(A) B19V (510⁹) was incubated with RBCs (0.5% in 100 μl PiBS) for 1h at pH 7.4 or 6.3. The cell suspensions were washed twice with a buffer of the same pH, except for one sample incubated at pH 6.3 and washed at pH 7.4. RBCs were incubated 30 additional minutes in the washing buffer at room temperature, washed twice, and viral DNA was extracted and quantified by qPCR. The results are presented as the mean ± SD of three independent experiments. *, p<0.05; **, p<0.01. (B) Visualization of the hemagglutination reaction by VLPs. RBCs were incubated with VLPs (510⁹) at pH 7.4 or 6.3 for 1h at room temperature. Prior to visualization by phase contrast microscopy, the samples were diluted with a buffer of the same pH or neutralized to pH 7.4. (C) Infectivity of Gb4-dissociated virus. Viruses bound to RBCs at pH 6.3 and dissociated at pH 7.4 were quantified by qPCR. Equal amounts of native and Gb4-dissociated virus were added to UT7/Epo WT or Gb4 KO cells. After 1h or 24h, cells were washed four times and NS1 mRNA was extracted and quantified by RT-qPCR. The results are presented as the mean ± SD of three independent experiments. ns, not significant.

Fig 6. Low pH-mediated interaction of B19V with Gb4 initiates active membrane processes.

(A) Quantification of B19V attachment to UT7/Epo WT and Gb4 KO cells at neutral and acidic pH. Cells were infected with B19V (10⁴ geq/cell) at 37°C for 1h followed by four washes. DNA was extracted and quantified by qPCR. The results are presented as the mean ± SD of three independent experiments. ***p<0.001; ns, not significant. (B) Detection of B19V capsids (860-55D) in UT7/Epo WT and Gb4 KO cells under neutral or acidic pH by IF. Cells were incubated for 30 min with functional (ΔC126) to block the VP1uR or non-functional (ΔN29) recombinant VP1u, as a control (S1 Fig) at 4°C. Subsequently, B19V (510⁴ geq/cell) was added for 1h at 37°C. (C) Gb4-mediated uptake of native B19V. Cells were preincubated with functional VP1u ΔC126 for 1h at 4°C prior to infection to block the VP1uR followed by incubation with B19V at 4°C or 37°C for 1h at pH 6.3. Non-internalized virus was removed by a washing step at neutral pH. Internalized viruses were detected by IF with antibody 860-55D against capsids. (D) Infectivity assay at neutral and acidic pH. WT cells were incubated for 30 min at 4°C with functional recombinant VP1u (ΔC126) to block the VP1uR or non-functional (ΔN29), as a control. Subsequently, B19V (510⁴ geq/cell) was added for 1h at 37°C in a buffer with the indicated pH. Cells were washed after 1h or further incubated for 24h at 37°C. NS1 mRNA was extracted and quantified by RT-qPCR. The results are presented as the mean ± SD of three independent experiments. ns, not significant. (E) Alternatively, 72h post-infection, capsid protein expression was examined by IF with antibody 3113-81C.

Fig 7. pH-mediated interaction of VP2-only particles with Gb4 triggers the uptake process independently of the VP1uR.

(A) Detection of VLPs bound to WT and Gb4 KO UT7/Epo cells at different pH values by Western blot using antibody 3113-81C. VLPs (10¹⁰) were incubated with cells (3x10⁵) at the indicated pH values for 1h at 37°C. GAPDH expression was used as a loading control. (B) Detection of VLPs (860-55D) in UT7/Epo WT and Gb4 KO cells under neutral or acidic pH by IF. (C) Gb4-mediated uptake of VLPs. Top and middle sections of cells incubated with VLPs at 4°C or 37°C for 1h at pH 6.3. (D) Gb4-mediated uptake of VLPs. Cells were incubated with VLPs at 4°C or 37°C for 1h at pH 6.3. Non-internalized virus was removed by a washing step at neutral pH. Internalized particles were detected by IF with antibody 860-55D against capsids.

Fig 8. Erythrocytes do not play a significant role as viral decoy targets during B19V viremia.

(A) Freshly collected blood samples (with EDTA as anticoagulant) and tested negative for antibodies against B19V, were spiked with B19V (10⁹ virions) and incubated for 1h at 37°C. The plasma and the RBC fractions were separated by centrifugation, the RBCs were washed with PBS (pH 7.4) and the viral DNA was extracted from both fractions and quantified by qPCR. (B) Detection of B19V in the RBC fraction by IF with an antibody against intact capsids (860-55D). (C) A component(s) in human plasma inhibits binding of B19V to RBCs. RBCs (0.5% in 100 μl PiBS) were incubated at pH 7.4 or 6.3 with B19V (5x10⁹) directly from an infected plasma sample or after purification by iodixanol density gradient centrifugation. After 1h at room temperature, the erythrocytes were washed four times with the corresponding buffer and viral DNA was extracted and quantified. The results are presented as the mean ± SD of three independent experiments. *, p<0.05; ***, p<0.001.