Plasmid construction

Cloning of EGFP-nsP3 was described previously [33] and cloned via Gateway technology into pcDNA/Dest40 and pIB-GW plasmid backbones for CMV and OpIE2 driven expression, respectively. Plasmids pIB-EGFP-nsP3.2, pIB-EGFP-nsP3-DDEL, and pIB-EGFP-nsP3-dUGA were generated by PCR from pIB-EGFP-nsP3 using the phosphorylated forward primer attB2-R-phos and reverse primers CHIKVnsP3-2R, CHIKVnsP3-DDEL-R, and CHIKVnsP3-dUGA-R, respectively. Plasmids pIB-EGFP-nsP3.7, pIB-EGFP-nsP3.8, and pIB-EGFP-nsP3.10 were generated from pIB-EGFP-nsP3 and pIB-EGFP-nsP3.2 using the phosphorylated reverse primer EGFP-R-phos and forward primers EcoRI-nsP3-161-F (nsP3.7 and nsP3.8), or EcoRI-nsP3-319-F (nsP3.10). Ae. albopictus Rin was amplified by RT-PCR (Invitrogen) from total RNA isolated from U4.4 cells using the Rin forward and reverse primers (Aalb-RIN-F and Aalb-Rin-R) and cloned into pGEM-T easy (Promega) and sequenced (Genbank accession number KP641128). To obtain pIB-Rin-mCherry, Rin was amplified by PCR from pGEM-Teasy-Rin using primers containing HindIII sites (HindIII-EcoRI-Aalb-Rin-F and HindIII-Aalb-Rin-R) and was inserted as a HindIII fragment into pIB-mCherry, in frame with and upstream of mCherry. Site-directed mutagenesis of pIB-EGFP-nsP3 and pIB-Rin-mCherry was performed using the forward and reverse primers fornsP3-P398A, nsP3-PPR401AAA, nsP3-FG479AA, nsP3-FG497AA, Rin-F34A and Rin-F34W. All constructs were verified by sequencing and primer sequences are listed in Table 1.

Transient expression of nsP3 and Rin

Insect cells were transfected with the indicated expression plasmids using Fectofly I (Polyplus) or ExpreS² Insect-TR (ExpreS2ion Biotechnologies). Mammalian cells were transfected with lipofectamine 2000 (Invitrogen). Twenty-four hours post transfection the fluorescence of EGFP-nsP3 and/or Rin-mCherry was analysed using a Zeiss Axio Observer Z1m inverted microscope in combination with an X-Cite 120 series lamp.

Rin Knockdown experiments

Linear DNA of Ae. albopictus Rin and firefly luciferase was generated by PCR from pGEM-T easy plasmids using the T7 universal primer (New England Biolabs) and T7-pGEM-Teasy-R and double-stranded (ds) RNA was synthesized in vitro with T7 RNA polymerase (Invitrogen). Knockdown in cell culture was performed by transfecting dsRNA into U4.4 cells grown in 24-wells plates (1 μg of RNA per well) using Fugene (promega). One day later, cells were transfected with plasmid pIB-EGFP-nsP3 to monitor nsP3-granule formation or infected with CHIKV at a multiplicity of infection (MOI) of five. At the indicated times post infection, the medium was removed from the cells and used in end point dilution assays on Vero E6 cells. The remaining cells were lysed in TRIzol (life technologies) reagent and total RNA was isolated. The RNA was DNase treated (Applied Biosystems) and reverse transcribed using random primers. Rin, S7 and genomic CHIKV cDNA were amplified (primers: Rin F2/R2, S7 F/R and nsP1 int F/R2) and detected with real-time PCR platinum SYBR Green (Invitrogen), in a Rotor Gene RG-3000 (Corbett Research).

In parallel experiments, cells were washed with PBS and lysed in SDS-loading buffer [100 mM Tris-Cl (pH 6.8), 4 % (w/v) sodium dodecyl sulfate (SDS), 0.2 % (w/v) bromophenol blue, 20 % (v/v) glycerol and 200 mM β-mercaptoethanol]. Samples were heated at 95 °C for 10 min, clarified by centrifugation for one min at 13 000 r.p.m and loaded on a 12 % SDS-Polyacrylamide gel. After electrophoresis, denatured proteins were transferred to an Immobilonmembrane (Millipore) for analysis by Western blotting. Membranes were blocked in 3 % skimmed milk in PBS with 0.05 % Tween 60 (PBST) for 1 h at room temperature. Membranes were washed three times for 5 min each with PBST and subsequently incubated for one hat room temperature with rabbit polyclonal anti-E2 (diluted 1 : 20000; [40]) and anti-β-tubulin (diluted 1 : 4000; Abcam) in PBST, respectively. Membranes were washed and treated with alkaline phosphatase conjugated with goat anti-rabbit IgG mAb (Sigma), diluted 1 : 3000 inPBST, for 45 min at room temperature. Membranes were washed twice for 5 min each with PBST and once for 10 min with AP buffer [100 mM NaCl, 5 mM MgCl2, 100 mM Tris/HCl (pH 9.5), 0.1 % Tween 20]. Proteins were detected by nitro blue tetrazolium chloride/BCIP staining (Roche).

In vivo knock down of Rin was performed in Ae. albopictus mosquitoes originating from la Reunion island (Providence, F11 generation). 500 ng of dsRin or dsLuc RNA was injected directly in the thorax of female mosquitoes (Drummond nanoject II). Two days post injection mosquitoes were either sacrificed and stored at -80 °C or orally infected with an infectious blood meal containing 10⁷ pfu/ml of CHIKV 06-021 strain. Mosquito rearing and preparation of the infectious blood meal was reported previously [36]. Fully engorged females were selected and incubated in climatic chambers (Binder) at 28 °C, with a light: dark cycle of 16 h: 8 h and 70 % relative humidity. Forced salivation were performed 6 days post-infection as described previously [35]. Saliva and mosquitoes were stored at -80 °C pending further analysis.

Infectivity assays

Frozen mosquitoes were dissected, separating bodies (abdomen and thorax) from the head. Individual mosquito bodies and heads were homogenized in the bullet blender storm (Next Advance) in 100 μl of DMEM Hepes (Gibco)-buffered medium supplemented with 10 % FBS containing penicillin (100 IU/ml), streptomycin (100 μg/ml), fungizone (2,5 μg/ml) and gentamycin (50 μg/ml) and spun down for 90 s at 14,000 rpm in a table top centrifuge. Thirty μl of the supernatant from the mosquito homogenate or the saliva-containing mixture was incubated on a monolayer of Vero cells in a 96-wells plate. After 2-4 h the medium was replaced by 100 μl of fresh cell culture medium, fully supplemented with antibiotics. Wells were scored for virus specific cytopathic effects (CPE) at three days post infection. Viral titres were determined using 10 μl of the supernatant from the mosquito homogenates in an end point dilution assay on Vero E6 cells. Infections were scored by CPE, three days post infection.

The conserved domain of CHIKV nsP3 is sufficient for multimerization but the variable domain is required for the formation of nsP3-granules

In mammalian cells, the C-terminal variable domain was found to be essential for nsP3 granule formation, and upon deletion the localization changed to a filamentous phenotype [33]. To determine which domains within nsP3 are responsible for the formation of nsP3-granules, truncated versions of nsP3 fused with EGFP (Fig. 1a) were expressed in insect cells (Fig. 1d). For these studies we used Sf21 cells because of their superior transfection efficiency as compared to mosquito cells. Removal of the entire C-terminal variable region (nsP3.2) resulted in the formation of filamentous, cytoplasmic structures. (Fig. 1d, left). To investigate whether the macrodomain could be eliminated from nsP3 without affecting its localization, it was deleted from EGFP-fused nsP3 and nsP3.2, yielding truncated mutants nsP3.7 and nsP3.8, respectively (Fig. ​(Fig.1a).1a). When expressed in insect cells, nsP3.7 showed an identical granular phenotype as full-length nsP3, whereas nsP3.8 formed filaments that were very similar to those produced by nsP3.2 (Fig. 1d, middle). We observed similar filaments upon expression of these nsP3.2 or nsP3.8 constructs in mammalian cells, and showed that they did not colocalize with cytoskeleton markers actin or tubulin [33]. To investigate if the C-terminal, variable region alone could cause granule formation, it was N-terminally fused to EGFP (nsP3.10) (Fig. 1a) and expressed in insect cells. The localization of nsP3.10 was diffuse, nuclear-cytoplasmic (Fig. 1d, right), showing that the C-terminal region of CHIKV nsP3 is required, but not sufficient for the formation of nsP3 granules.

NsP3 granules co-localize with Rasputin, the insect homolog of mammalian G3BP

In SINV nsP3 pull-down studies, G3BP and its insect homologue Rasputin (Rin) were identified as predominant nsP3-interacting proteins in virus-infected mammalian and mosquito cells, respectively [28]. To study a putative interaction of CHIKV nsP3 with mosquito Rin in live cells, the gene encoding Ae. albopictus Rin was amplified by RT-PCR from total RNA isolated from U4.4 cells using PCR primers specific for Ae. aegypti Rin. We cloned and sequenced the obtained amplicon. The Ae. albopictus Rin sequence (Genbank:KP641128) was 307 amino acids longer than human G3BP1, but the nuclear transport factor 2 (NTF2)-like domain, the RNA recognition motif (RRM) and the arginine glycine-rich (RGG-) box were conserved between these species (Fig. 2a).

Open in a separate window

Fig. 2

CHIKV nsP3 sequesters mosquito Rasputin into cytoplasmic granules. a Schematic representation of Ae.albopictus-Rin-mCherry. NTF2; nuclear transport factor 2-like domain, RRM: RNA recognition motif, RGG; arginine-glycine rich box. b Ae.albopictus-Rin-mCherry or EGFP-nsP3 were transiently expressed in SF21 insect cells and display diffuse and granular intracellular distributions, respectively. c Ae.albopictus-Rin-mCherry and EGFP-nsP3 (top) or EGFP-nsP3.8 (bottom) were co-transfected into insect cells. Magnified images are presented

The subcellular localization of Rin was studied by transient expression in insect cells as C-terminal fusion with mCherry, in a similar fashion to a previously described and functional G3BP-EGFP fusion [31]. When expressed in insect cells, Rin was evenly distributed throughout the cytoplasm (Fig.​(Fig.2b,2b, left). However, when Rin was co-expressed with nsP3, which localized to small nsP3 granules (Fig. 2b, right), both proteins displayed strong co-localization and formed much larger granules (Fig. 2c, top). These nsP3- and Rin-positive granules or aggregates were larger and more asymmetrical than normal nsP3-granules. In this experiment, Rin was transiently (over) expressed, which may explain the large size of the granules. In contrast, when mCherry-Rin was co-expressed with the C-terminal truncated, filamentous mutants EGFP-nsP3.8 (Fig. 2c, bottom) or EGF-nsP3.2 (not shown), Rin did not co-localize with the filaments formed by these mutants and retained its diffuse, cytoplasmic localization (Fig. 2c, bottom). In conclusion, the C-terminal hypervariable domain of nsP3 is important for the interaction with Rin in insect cells.

The C-terminal TFGD repeats in the variable domain of CHIKV nsP3 interact with Rasputin

Previously we showed that transiently expressed CHIKV nsP3 sequesters G3BP into cytoplasmic granules in mammalian cells. Deletion of a conserved SH3-domain binding motif (PVAPPRRRR) in the variable domain of nsP3 resulted in a diffuse nuclear/cytoplasmic localization of nsP3 and abrogated the interaction between nsP3 and G3BP, restoring the potential of the cell to respond to oxidative stress [33]. A recent study, however, convincingly showed that the interaction between nsP3 and mammalian G3BP depends on two conserved repeats (with core sequence TFGD) in the variable domain of nsP3 [43]. To further investigate which domain(s) of nsP3 are crucial for the interaction between CHIKV nsP3 and Rin, both the SH3-domain binding motif and conserved TFGD repeats in the variable domain of nsP3 were mutated using site-directed mutagenesis (Fig. 3a). When EGFP-nsP3 and Rin-mCherry were transiently expressed in insect cells, both wild type proteins displayed perfect co-localization. Deletion of the SH3-domain binding motif (nsP3-d398/406) resulted in a diffuse nuclear/cytoplasmic distribution of nsP3 and a diffuse mainly cytoplasmic localization of Rin (Fig. 3b, second panel), identical to the distribution of nsP3 and G3BP in mammalian cells [33]. However, when conserved proline and/or argenine residues from the SH3-domain binding motif were substituted for alanines (nsP3-P398A and nsP3-PPR401AAA) the co-localization of nsP3 and Rin in cytoplasmic nsP3-granules remained unchanged (Fig. 3b).

Open in a separate window

Fig. 3

The C-terminal TFGD repeats of nsP3 interact with mosquito Rasputin. a Schematic representation of nsP3 and the conserved domains within the C-terminal variable domain. Deletions and mutations are indicated in the amino acid sequence. Insect cells (Sf21) were co-transfected with Rin-mCherry and either one of the EGFP-nsP3 variants. b SH3-domain binding motif, nsP3-d398/406, nsP3-P398A or nsP3-PPR401AAA. c C-terminal repeats, EGFP-nsP3-FG479AA, EGFP-nsP3-FG497AA or EGFP-nsP3-FG479AA/FG497AA

Next, we mutated both domains of the two conserved TFGD repeats separately or together resulting in pIB-EGFP-nsP3-FG479AA, pIB-EGFP-nsP3-FG497AA, and pIB-EGFP-nsP3-FG479AA/FG497AA. The single and double EGFP-nsP3 TFGD mutants were transiently expressed in insect cells together with Rin-mCherry. Both the single FG479AA and FG497AA mutants still sequestered Rin into nsP3-granules (Fig. 3c, top and middle panels). The double TFGD mutant, however, displayed a completely diffuse intracellular distribution of Rin (Fig. 3c, bottom panel) but retained a normal granular distribution similar to wildtype EGFP-nsP3 (Fig. 1b).

These results indicate that deletion of the SH3-domain binding motif does abrogate the formation of nsP3 granules, but the formation of nsP3-granules and the interaction with Rin is retained when conserved amino acids within this domain are substituted for alanines. However, the formation of CHIKV nsP3-Rin-granules is also abrogated when both the C-terminal conserved TFGD repeats are mutated, suggesting that these motifs are involved in the nsP3-Rin interaction.

The NTF2-like domain of mosquito Rasputin interacts with CHIKV nsP3

The NTF2-like domain of G3BP has been shown to interact with SINV nsP3 in mammalian cells [29]. The first 140 amino acids of mosquito Rin show high sequence homology to that of NTF2-like domains including the NTF2-like domain of human G3BP (Fig. 4a). Protein structure prediction of Ae. albopictus Rin revealed a 100 % structural homology of the NTF2-like domain with those of G3BP and Drosophila Rin (Fig. 4b, left), which contains a binding pocket for FxFG containing peptides [44, 45]. Here we mutated a phenylalanine of Rin (F34) that is expected to interact with the phenylalanines of FxFG domains via pi-stacking (Fig. 4b, right) [44]. Indeed, amino acid substitutions in Rin (F34W or F34A), within PiB-A.alb-Rin-mCherry (Fig. 2a) strongly reduced the sequestration of Rin into nsP3-granules (Fig. 4c), indicating that this amino acid is essential for nsP3-Rin interaction.

Open in a separate window

Fig. 4

nsP3 interacts with the NTF2-like domain of mosquito Rasputin. a Protein alignment of the Rin/G3BP NTF2-like domains from Ae. albopictus (KP641128), Ae. aegypti, (XP_001651045), Culex quincefasciatus (XP_001861860), Anopheles gambiae (XP_001688309), Anopheles sinensis (KFB40464), Drosophila melanogaster (AF231031), Ixodes ricinus (GANP01009274), and Homo sapiens (CAG38772). Genbank accession numbers in brackets. Alignment made with CLUSTALX and modified using Genedoc. b Structural modeling of Ae. albopictus-Rin modeled onto D. melanogaster Rin. Modeling was performed using the Phyre2 server (www.sbg.bio.ic.ac.uk/phyre2), results were visualized with PyMOL (www.pymol.org). Left, Ae. albopictus Rin (deep blue) is depicted together with D. melanogaster Rin (cyan). Right, the FxFG binding pocket is shown and phenylalanine 34 is highlighted in red. c Sf21 insect cells were co-transfected with EGFP-nsP3 and either Rin-mCherry (wild type), Rin-mCherry F34A or Rin-mCherry F34W

Effect of Rasputin silencing on the formation of CHIKV nsP3-granules

So far, we have shown that nsP3 has a granular localization and that Rin has a diffuse cytoplasmic localization when expressed individually. To investigate whether the formation of nsP3-granules requires Rin, localization of EGFP-nsP3 was studied after Rin expression was silenced in mosquito cells using dsRNA-mediated RNAi. U4.4 mosquito cells were transfected with dsRNA from either Rin (dsRin) or firefly luciferase (dsLuc) as a control. Semi-quantitative RT-PCR on Ae. albopictus Rin mRNA, normalized by housekeeping gene S7 mRNA, showed a 90 % reduction in Rin mRNA when cells were transfected with dsRin RNA (Fig. 5a). Subsequent transient expression of EGFP-nsP3 displayed clear nsP3-granules in both dsRin and dsLuc transfected cells (Fig. 5b). This result indicates that the formation of CHIKV nsP3-granules in mosquito cells is independent of (high levels of) mosquito Rasputin.

Open in a separate window

Fig. 5

Formation of nsP3-granules is independent of mosquito Rasputin. a. Rin silencing was determined by semi-quantitative RT-PCR on Ae. albopictus Rin mRNA, relative for the internal control S7. Bars represent relative Rin mRNA expression normalized to dsLuc transfected samples. Error bars indicate standard deviation of duplicate samples from a single representative experiment, which is presented in B. b. U4.4 mosquito cells were transfected with dsRNA against Ae. albopictus Rin or luciferase. Twenty-four hours later these cells were transfected with wild type EGFP-nsP3, displaying a granular localization

We thank Els Roode for technical support concerning the BSL3 laboratory in Wageningen. We acknowledge Marie Vazeille for injecting mosquitoes and Anubis Vega Rua and Laurence Mousson for their assistance in forced salivation.