Alphavirus nsP3 organizes into tubular scaffolds essential for infection and the cytoplasmic granule architecture

Abstract

Alphaviruses, such as chikungunya virus (CHIKV), are mosquito-borne viruses that represent a significant threat to human health due to the current context of global warming. Efficient alphavirus infection relies on the activity of the non-structural protein 3 (nsP3), a puzzling multifunctional molecule whose role in infection remains largely unknown. NsP3 is a component of the plasma membrane-bound viral RNA replication complex (vRC) essential for RNA amplification and is also found in large cytoplasmic aggregates of unknown function_. Here, we report the cryo-electron microscopy (cryo-EM) structure of the CHIKV nsP3 at 2.35 Å resolution. We show that nsP3 assembles into tubular structures made by a helical arrangement of its alphavirus unique domain (AUD). The nsP3 helical scaffolds are consistent with crown structures found on tomographic reconstructions of the mature viral RCs. In addition, nsP3 helices assemble into cytoplasmic granules organized in a network of tubular structures that contain viral genomic RNA and capsid as well as host factors required for productive infection. Structure-guided mutagenesis identified residues that prevent or disturb nsP3 assemblies, resulting in impaired viral replication or transcription. Altogether, our results reveal an unexpected nsP3-dependent molecular organization essential for different phases of alphavirus infection.

Fig. 1. Structure of CHIKV nsP3 helical scaffolds (HSs).

a Alphavirus genomic organization and nsP3 domains. b Cryo-EM map of nsP3 HSs at 2.52 Å resolution (EMD: 17729) is represented in light gray. The left panel shows a longitudinal view of the scaffold with two turns of the two-start helix of AUD domains (hA and hB) colored in light and dark green, respectively. The N and C terminus of each protomer are colored in blue and red, respectively. The right panel shows a transversal section of the same structure, 11 protomers per turn are visible, the dimensions of the internal and external diameter of the AUD scaffold are indicated. c same representation of the cryo-EM map calculated with a larger mask at 2.98 Å resolution (EMBD: 17730) colored in dark gray. Additional densities corresponding to the MD are visible and colored in cyan and blue for the hA and hB, respectively. The right panel shows the transversal section of the structure as in (a). The outer diameter of the scaffold including the MDs is indicated. d Cryo-EM map after local 3D classification at 6.4 Å resolution (EMBD: 17730) is shown in yellow. The AUDs are colored in light green and labeled. The N and C terminus of the AUD are colored in blue and red respectively. This map shows the connection of the AUD N terminus with the density corresponding to the MD, and also the interaction of the beginning of the HVD at the C terminus of the nearby AUD n + 1 contacting the protomer n MD. e Cartoon representation of the AUD domain with the oligomerization modules (O-modules) which mediate lateral and axial contacts colored in light gray and dark green, respectively, and labeled. The oligomerization loop, which mediates contacts in both axial and lateral interface, is colored in cyan and labeled. f The interprotomer interactions are shown in the context of the helical scaffold. The two helices of AUD, hA and hB, are represented in light and dark cartoons and labeled. The AUD An is colored in green and the residues mediating interprotomer contacts are shown in spheres and colored as in panel d. The residues mediating contacts from Bn and A n - 1 protomers are also shown in spheres. The axial (Bn/An) and lateral (An/An-1) interfaces are indicated by dashed line squares. g Detailed representation of axial and lateral contacts between AUDs are shown in the upper and bottom panels, respectively. AUDs are labeled, represented in cartoon and colored as in panel (e). The residues mediating interprotomer contacts are shown in sticks and labeled. Hydrogen bonds are shown by yellow dashed lines.

Fig. 2. Analysis of charge distribution on the AUD helical scaffolds.

a Surface charges representation of two sides of the AUD monomer. The monomer on the left has the same orientation than the monomer in Fig. 1e. The O-loop displays negative surface charges while the axial O-module, partly including the Zn finger, displays positive charges. b A section through the diameter of the scaffold shows the charge distribution of the interior which is rather neutral, equally combining positive and negative charges. The right panel shows the surface of the positively charged surface of the tubes. The positively charged interstitial pores of the helices connect with the external positive charged surface suggesting that naked RNA could transit the pores and bind the scaffold surface. c Same representation of the bottom and top of the HSs with positively and negatively charges, respectively, showing the electrostatic nature of the AUD oligomerization and helix formation. The surface electrostatics were calculated using APBS (https://www.poissonboltzmann.org/) on the ranges indicated by the scale-bars for panel (a) and (b, c) respectively.

Fig. 3. Blockage of the AUD oligomerization interfaces in the non-cleaved nsP3 polyprotein.

a Superposition of the Alphafold2 predicted structure of the uncleaved nsP2-3 on the nsP3 HSs. One turn of the helical scaffold is shown in dark and light gray cartoon (for hA and hB respectively). The model is superposed in the hA AUD of the helix. It is shown with nsP2 protease surface colored in magenta, the MTase domain in olive green, the nsP3 MD domain in blue and the AUD in green cartoons. The protease domain completely blocks the axial contacts of the helical scaffold in the nsP2-3 uncleaved protein structure. b Same structure than in (a), showing only the protease domain and the AUD domain. On the left panel the residues mediating axial interactions shown in Fig. 1g are shown in spheres and colored according to panel a. On the right panel we have the same representation for the lateral contacts. The residues mutated in the replicon experiments (Y200A, P247A/V248A and K302A) are colored in red and labeled (see section “Mutations on the nsP3 AUD oligomerization interfaces affect RNA synthesis and production of structural proteins”). Only the Y200A is in the interface between the nsP2 protease and the AUD. The N terminus of the AUD connecting with the macrodomain is colored in blue.

Fig. 4. Superposition of nsP3 HSs on tomographic reconstructions of membrane-associated CHIKV RCs.

a Superposition of nsP1 capping pores and the nsP3 scaffolds on the tomographic reconstruction of CHIKV RCs. Top and lateral views are shown in the top and bottom panels respectively. The volume is shown as a transparent gray surface. nsP1 and nsP3 AUDs are colored by protomer and shown in spheres, the MDs of nsP3 are shown in cartoon and colored in cyan. Their position with respect to the AUDs was determined based on their fitting on the map shown in Fig. 1b (EMD: 17730). The tomographic volume was calculated applying a 12-fold symmetry and consequently has 12 outer volumes on the region which are shown numbered. The superposition shows the consistent matching in size and shape of AUD helices with the rings above the nsP1 capping pores and the outer spherical volumes have a partial matching with the MDs positions found in HSs. However, the structure of the nsP3 helical scaffold have 11,37 protomers per turn instead of 12, slightly breaking the symmetry with the nsP1 pores and the volume reconstructions. Superposition allows to confidently assign the central ring above nsP1 to the AUDs and the peripheral volumes to the MDs. b same representation of the superposition of capping pores and nsP3 scaffolds on the volume of RCs reported by Tan et al.. The volume reaches higher resolution and only shows electron density for the AUD region of the nsP3 scaffold. The volume was also calculated imposing C-12 symmetry.

Fig. 5. NsP3 HS formation requires nsP2-nsp3 precursor cleavage.

a–c U2OS cells were transfected with a plasmid encoding the polyprotein precursor (P1234) or cleavage-deficient mutants (^ indicates the uncleavable sites). a Cell lysates were separated by SDS-PAGE and non-structural proteins processing was ascertained by immunoblot analysis with nsP1, nsP2, nsP3, nsP4 or GAPDH antibodies. b The subcellular localization of nsP1, nsP2 and nsP3 was assessed by immunofluorescence at 48 h post-transfection. a, b Data shown are representative of 2 independent experiments. c Graph bar showing the number of nsP3 aggregates per cell (left) and the average aggregate size (right) from two independent biological replicates (n = 20 cells in mock, n = 43 cells for P1234, n = 42 cells in P1^234, n = 21 cells in P12 ^ 34 and n = 31 cells in P1 ^ 2 ^ 34). Data are presented as mean value ± SD and adjusted P-values are calculated by one-way ANOVA with Kruskal-Wallis multiple comparisons test. Significant differences denoted by ****P < 0.0001, * P = 0.0232. Source data are provided in the Source Data file.

Fig. 6. CHIKV nsP3 assembles into tubular structures that define the architecture of alpha-granules.

a Transmission electron microscopy of human fibroblasts infected with CHIKV 21 strain (MOI of 10) for 24 h (upper panel) or stably expressing the CHIKV nsP3 (lower panel). Longitudinal (left) and transversal (right) views of the nsP3 tubes are shown. Images are representative of two independent experiments. Scale bars, 500 nm (mock panels) and 200 nm (infection and nsP3-overexpression panels). b Transmission electron microscopy of human fibroblasts transfected with empty plasmid (mock) or MAYV, OONV or EEEV nsP3 plasmids. Scale bars, 500 nm. c The left panel shows a micrograph corresponding to a section of alpha granules perpendicularly cutting a bundle of HSs. A 0.2 µm² is depicted. 16 HSs are present and numbered within the 0.2 µm² square. The scaffold numbered as 13 is superposed to three concentric circles of the dimensions obtained for the HSs (see Fig. 1) including the internal channel (green circle), the MDs (blue circle), the and external density around the tube (red circle). Circles diameter and area were calculated with ImageJ for each one of the 16 HSs in the image (bottom panels). Scale bar, 20 nm. Data are presented as mean ± SD from one experiment. Source data are provided as a Source Data file. d Human fibroblasts were infected with CHIKV 21 strain for 24 h. Cells were immunostained with an anti-Capsid or anti-nsP3 specific Ab. Viral genomic RNA was detected using a CHIKV-genome specific fluorescent RNA probes. Co-localization analysis was performed using JACoP plugin implemented in ImageJ. Scale bars, 20 µm. Images are representative of 2 independent experiments.

Fig. 7. nsP3 helical scaffolds are required for CHIKV infection.

a–c Characterization of nsP3 AUD mutants. U2OS cells were transfected with plasmids encoding FLAG-tagged wild-type CHIKV nsP3 or mutants harboring substitutions of amino acids involved in AUD-AUD contacts (Y200A, P247A/V248A, and K302A/V303A). a At 48 h after transfection, cells were fixed and nsP3 sub-cellular localization was assessed by immunofluorescence using an anti-FLAG mAb. Images were acquired by confocal microscopy and are representative of 2 independent experiments. White arrows indicate the presence of nsP3 aggregates. Scale bars, 20 μm. b The number of nsP3 aggregates per cell was assessed using ImageJ. Data are presented as min to max with all data points shown (n = 20 cells for each condition). Adjusted P-values are calculated by one-way ANOVA with Dunnett’s multiple comparison test (****P < 0.0001). c Negative staining micrographs of nsP3 in vitro assembly experiments after incubation at 0.1 M NaCl for 1 h showing the assembly of HSs. White scale bars are 100 nm long. a–c Data are representative of 2 independent experiments. d, e HEK293T (d) or U2OS (e) cells were transfected with the indicated in vitro-transcribed and capped CHIKV RNA. Viral titers in the supernatants were quantified by plaque assays 48 h after RNA transfection. f HEK293T were transfected with the indicated in vitro-transcribed and capped SINV RNA. Viral titers in the supernatants were quantified by plaque assays 48 h after RNA transfection. Data in (d–f) are presented as min to max with all data points shown (n = 3 independent experiments performed in duplicate). Adjusted P-values are calculated by one-way ANOVA with Dunnett’s multiple comparison test (***P = 0.0003). g 293 T were transfected with CHIKV-D-Luc-SGR wild type (WT) and the indicated mutant RNA and harvested for both RLuc and Fluc assays at the indicated times. GAA: inactive mutant of nsP4 polymerase. Data are presented as mean value ± SD (n = 3 independent experiments performed in quadruplicate) and adjusted P-values are calculated by two-way ANOVA with a Dunnett’s multiple comparison test (****P < 0.0001). Source data are provided as a Source Data file.

Fig. 8. Model of the role of nsP3 HSs during the alphavirus infection.

1 After vRNA release in the cytoplasm, the P1234 polyprotein is processed into P123 + 4, forming the early vRC and spherules that house the dsRNA intermediate. Axial contacts between adjacent AUD promoters trigger the formation of a helical scaffold (HSs) required for dsRNA synthesis. 2 A conformational change switch concomitant to the processing of the P123 by nsP2 releases the AUD loop and mediates nsP3 oligomerization and subsequent cytoplasmic crown formation, leading to the mature vRC synthesizing positive gRNAs and sgRNAs. 3 As infection progresses, nsP3 accumulates into alpha-granules organized by highly-ordered networks of helical tubular scaffolds. 4 Because these structures contain vRNA and capsid, nsP3 assemblies may sort the vRNA and deliver it in the cytoplasm for structural protein synthesis and/or nucleocapsid assembly. 5 Once the nucleocapsid reaches the plasma-membrane, the capsid/E2/E1 envelop protein interaction triggers an organized budding event leading to the release of infectious viral particles.