Dengue and Zika Virus Capsid Proteins Contain a Common PEX19-Binding Motif

Abstract

Flaviviruses such as dengue virus (DENV) and Zika virus (ZIKV) have evolved sophisticated mechanisms to suppress the host immune system. For instance, flavivirus infections were found to sabotage peroxisomes, organelles with an important role in innate immunity. The current model suggests that the capsid (C) proteins of DENV and ZIKV downregulate peroxisomes, ultimately resulting in reduced production of interferons by interacting with the host protein PEX19, a crucial chaperone in peroxisomal biogenesis. Here, we aimed to explore the importance of peroxisomes and the role of C interaction with PEX19 in the flavivirus life cycle. By infecting cells lacking peroxisomes we show that this organelle is required for optimal DENV replication. Moreover, we demonstrate that DENV and ZIKV C bind PEX19 through a conserved PEX19-binding motif, which is also commonly found in cellular peroxisomal membrane proteins (PMPs). However, in contrast to PMPs, this interaction does not result in the targeting of C to peroxisomes. Furthermore, we show that the presence of C results in peroxisome loss due to impaired peroxisomal biogenesis, which appears to occur by a PEX19-independent mechanism. Hence, these findings challenge the current model of how flavivirus C might downregulate peroxisomal abundance and suggest a yet unknown role of peroxisomes in flavivirus biology.

Keywords: Zika virus; capsid protein; dengue virus; flavivirus; human peroxin PEX19; peroxisomes.

Figure 1

Peroxisomes are required for optimal DENV replication. (A) HFFF cells were infected with DENV-2 at an MOI of 2 or 0.02 and the appearance of infectious virus in the extracellular supernatant quantified by plaque assay over a 72-h time course of infection. (B) Primary human fibroblasts (HFFF) and PEX19-deficient (∆PEX19) or PEX3-deficient (∆PEX3) patient-derived fibroblasts were infected with DENV-2 at an MOI of 2 or 0.02 and extracellular infectious virus quantified by plaque assay at the indicated time points. Graphs show mean +/− standard deviation. pfu, plaque-forming units. *** p < 0.001 (two-way ANOVA, Tukey correction for multiple comparisons).

Figure 2

DENV C protein binds PEX19 via a PEX19-binding motif. (A) ^SBPPEX19 was co-expressed in HEK293T cells together with ALD^flag WT or with ALD^flag harbouring mutations R69A, V70G, R80A, and L81G within its PEX19-binding motif (Mut^.). ^SBPPEX19-DENV C^flag complexes were affinity purified using streptavidin agarose bead and analyzed by immunoblotting using PEX19- or flag-specific antibodies, respectively. (B) Sequence alignment of ZIKV and DENV C proteins. The PEX19 binding motif is boxed in red, and the four alpha helices (a1–a4) [32] are indicated by grey boxes. (C) ^SBPPEX19 was co-expressed in HEK293T cells together with the DENV C WT construct or DENV C with mutations within the predicted PEX19-binding motif (KL45, RF55, or both). ^SBPPEX19-DENV C^flag complexes were affinity purified using streptavidin agarose beads and analyzed by immunoblotting using PEX19- or flag-specific antibodies, respectively. (D) ALD^flag, the DENV C^Flag WT, or mutant version (KL45/RF55) were expressed in HEK293T cells. Protein complexes were affinity purified using PEX19-specific antibodies. An unrelated antibody was used as negative control (control). Samples were subsequently analyzed by immunoblotting using PEX19- or flag-specific antibodies, respectively. The amount loaded for input represents 15% of the amount loaded for the IP.

Figure 3

ZIKV capsid protein binds PEX19 via a conserved PEX19-binding motif. (A,C) ^SBPPEX19 was co-expressed in HEK293T cells together with the ZIKV C^Flag WT construct or versions with mutations (RM45, RF55, RM45/RF55, RV23, or KR31), as indicated. ^SBPPEX19-ZIKV C^flag complexes were affinity purified using streptavidin agarose bead and analyzed by immunoblotting using PEX19- or flag-specific antibodies, respectively. (B) The ZIKV C WT, or the C variant RM45/RF55 was expressed in HEK293T cells. Protein complexes were affinity purified with PEX19-specific antibodies. An unrelated antibody was used as negative control for the co-IP (control). Samples were subsequently analyzed by immunoblotting using PEX19- or flag-specific antibodies, respectively. An aliquot (15%) of the input is shown. The amount loaded for input represents 15% of the amount loaded for the IP for all western blots.

Figure 4

PEX19 is not involved in C transport to organelles. (A) A549 cells were transfected with plasmids encoding for the human peroxisomal membrane marker ALD (ALD^flag), the HSV-1 protein ICP47 [35] (ICP47^flag), and the WT or mutated variants of the DENV (DENV C^flag KL45/RF55) and ZIKV (ZIKV C^flag RM45/RF55) C proteins. 24 h post-transfection, cells were fixed and stained for indirect immunofluorescence using catalase-specific and flag-specific antibodies. (B) Colocalization analysis of flag-positive structures and catalase using Pearson’s coefficient. Statistical analysis was carried out using an ordinary one-way ANOVA with Sidak’s multiple comparisons test (n > 7; ns = non-significant). (C) A549 cells were transfected with pIRES2-EGFP derived plasmids encoding for WT or mutated variants of the ZIKV or DENV C proteins (ZIKV C^flag WT, DENV C^flag WT, ZIKV C^flag RM45/RF55 or DENV C^flag KL45/RF55, respectively). At 48 h post-transfection, cells were processed for indirect immunofluorescence using anti-flag antibody for the detection of C proteins (red). Cells were also stained with LipidTOX deep red for lipid droplets detection (indicated in green). Nuclei were stained by DAPI and overlaid with anti-mouse 594 and LipidTOX deep red signals (merge). Intrinsic EGFP fluorescent is shown in gray. (D) Colocalization analysis of flag tag with lipid droplets using Pearson’s coefficient. Statistical analysis was carried out using an ordinary one-way ANOVA with Sidak’s multiple comparisons test (n > 9; ns = non-significant, * = p < 0.05).

Figure 5

Expression of ZIKV capsid proteins causes loss of peroxisomes. (A) A549 cells were transfected with pIRES2-EGFP-derived plasmids encoding the ZIKV C WT or mutated variant (ZIKV C RM45/RF55) of the ZIKV C protein, or the empty vector as negative control. At 72 h post transfection, cells were processed for indirect immunofluorescence using mouse anti-PMP70 antibodies for the detection of peroxisomes (in red). Images show the top and side view Z stack reconstruction of a single cell. (B) Z-stack images of GFP-positive cells were acquired in 0.2 μm intervals, and peroxisome abundance analysis was performed using the total volume of peroxisomes:total volume of the cell ratio. Statistical analysis was carried out using an ordinary one-way ANOVA with Tukey’s multiple comparisons test (n > 20; ns = non-significant, ** = p < 0.01, **** = p < 0.0001).

Figure 6

ZIKV capsid impairs peroxisome biogenesis. (A) Time course of peroxisomal rescue in PEX19-deficient fibroblasts. PEX19-deficient patient-derived human fibroblast were transfected with the plasmid pNoGFP-EGFP-SBP-PEX19 (EGFP-PEX19) to rescue peroxisomal biogenesis. The co-expressed mCherry-SKL serves as a peroxisomal lumenal marker. Cells were fixed and analyzed by intrinsic EGFP and mCherry fluorescence either 24 h or 48 h post-transfection. Nuclei were stained using DAPI (blue). (B) Peroxisomal rescue is induced by PEX19. PEX19-deficient fibroblasts were transfected with the plasmid pIRES2-mCherry-SKL/PEX19 or pIRES2-mCherry-SKL/PEX3. In parallel, cells were co-transfected with the empty pIRES2-EGFP vector. At 48 h post-transfection, cells were fixed and analyzed for intrinsic GFP and mCherry fluorescence. (C) PEX19-deficient patient-derived human fibroblast were transfected with the plasmid pIRES2-mCherry-SKL/PEX19 to rescue peroxisomal biogenesis by the human PEX19 WT cDNA, as well as the peroxisomal lumenal marker protein mCherry-SKL. In parallel, cells were co-transfected with pIRES2-EGFP derived plasmids encoding the WT or mutated variant of the ZIKV C protein (ZIKV C RM45/RF55), or the empty pIRES2-EGFP plasmid as negative control. At 48 h post-transfection, cells were fixed and analyzed for intrinsic GFP and mCherry fluorescence. (D) Peroxisomal abundance in GFP-positive cell was quantified using ImageJ. Statistical analysis was carried out using an ordinary one-way ANOVA with Tukey’s multiple comparisons test (n > 35; ns = non-significant, **** = p < 0.0001).