SNAP-tagged Chikungunya Virus Replicons Improve Visualisation of Non-Structural Protein 3 by Fluorescence Microscopy.

Chikungunya virus (CHIKV), a mosquito-borne alphavirus, causes febrile disease, muscle and joint pain, which can become chronic in some individuals. The non-structural protein 3 (nsP3) plays essential roles during infection, but a complete understanding of its function is lacking. Here we used a microscopy-based approach to image CHIKV nsP3 inside human cells. The SNAP system consists of a self-labelling enzyme tag, which catalyses the covalent linking of exogenously supplemented synthetic ligands. Genetic insertion of this tag resulted in viable replicons and specific labelling while preserving the effect of nsP3 on stress granule responses and co-localisation with GTPase Activating Protein (SH3 domain) Binding Proteins (G3BPs). With sub-diffraction, three-dimensional, optical imaging, we visualised nsP3-positive structures with variable density and morphology, including high-density rod-like structures, large spherical granules, and small, low-density structures. Next, we confirmed the utility of the SNAP-tag for studying protein turnover by pulse-chase labelling. We also revealed an association of nsP3 with cellular lipid droplets and examined the spatial relationships between nsP3 and the non-structural protein 1 (nsP1). Together, our study provides a sensitive, specific, and versatile system for fundamental research into the individual functions of a viral non-structural protein during infection with a medically important arthropod-borne virus (arbovirus).


Formation of filamentous nsP3 structures during alphavirus replication.
Truncation of ten C-terminal amino acid residues from SFV-encoded nsP3 and expression of the mutant protein in    Parasit Vectors 8: 464, 2015). However, the nsP3-positive filaments described in the above studies differ from the nsP3-positive rods described in this study, since SNAP-nsP3 rods rarely formed long filaments as seen with truncation mutants.

Combination of Airyscan microscope and 3-D surface rendering. Airyscan confocal
super-resolution microscopy was used to evaluate three-dimensional morphological features below the diffraction limit of light using a straightforward work flow. The optical sectioning capabilities of this system allowed us to acquire high-quality series of optical slices and reconstruct 3-D volumes from these image stacks. We found that the entire depth of fixed HuH-7 cells could be captured by a series of 30-40 images, with individual acquisition times of 5-20 s for each image, depending on the number of channels imaged. The resulting images of SNAP-labeled nsP3 were high in contrast and could be processed within minutes using a commercial software for 3-D image processing and analysis. Visualisation of these structures by fluorescence microscopy can be technically challenging due to the large difference in fluorescence intensity between high-density clusters, such as fiber-like structures and low-density clusters, such as small granules. Using the thresholding functions of 3-D processing software, we have been able to visualise the surfaces of nsP3-positive clusters that have both high and low protein density.

Links between lipid metabolism and alphavirus replication. Previously, treatment
with inhibitors of various lipid biosynthesis pathways reduced RNA replication of Semliki
Note that Mitotracker Green FM (Thermo Fisher Scientific) was used as a cellular counterstain for long-term tracking of live cells by confocal microscopy.
Immunofluorescence assays for standard confocal microscopy. At the indicated times (12 h, 28 h post-transfection) cells were washed with PBS and fixed with 4% formaldehyde in PBS. For analysis of stress granule markers or staining of dsRNA, cells were then permeabilised by incubating with 100% methanol at -20°C. Cells were then blocked for 30-60 min in a buffer containing 10% fetal calf serum in PBS. For staining of nsP1 and nsP3, cells were permeabilised by also including 0.1% Triton X-100 in this blocking buffer. Cells were then washed with PBS and stained with the aforementioned primary antibodies overnight in an antibody dilution buffer containing 5% bovine serum albumin in PBS (with dilutions of 1:1000 for nsP1 and nsP3, 1:200 for J2, 1:100 for stress granule markers). Triton X-100 was added at 0.1% to the antibody dilution buffer that was added to cells that had already been permeabilised with Triton X-100, but left out in cells that had been permeabilised with methanol to avoid additional extraction of proteins in methanol-permeabilised cells. Primary antibody incubation was followed by a minimum of three washes in PBS and a 1-hour-incubation with dye-conjugated secondary antibodies in antibody dilution buffer. Where indicated, cells were counterstained with fluorescent dyes for nuclei (DAPI), lipid droplets (monodansylpentane), or α-mannopyranosyl and α-glucopyranosyl residues (concanavalin A), which are found in mammalian cell membranes, as described in Supplementary Text S1. Coverslips were mounted onto glass slides by the addition of ProLong Diamond Antifade Mountant (Thermo Fisher Scientific).

Quantification of replicon-driven ZsGreen-expression.
Lipofection was used to transfect cells, which had been seeded in triplicate into 24-wells, with ZsGreencontaining replicons CHIKV repl sg-ZsGreen , CHIKV repl SNAP-P3 sg-ZsGreen , and CHIKV repl mCherry-P3 sg-ZsGreen . Untransfected cells were included as a negative control. For live-cell studies, we captured microscopy images at hourly intervals for a total time of 20h starting at 7h post-transfection. Images were acquired using an IncuCyte ZOOM system (Essen BioScience), which consists of an automated phase-contrast and fluorescent microscope housed within a humidifying incubator connected to a 5% CO2 line. In each well, the system was set to take nine images with the following software settings: Nikon 10x objective, dual-colour-filter module (model 4459), two image channels (Green and Phase with 1392 x 1040 pixels at 1.22 µm per pixel, acquisition time of 400 ms and 1000 ms respectively). Various metrics including overall cell confluency, green object count (per mm 2 ), green object confluence (in %), mean fluorescence intensity (the green object's mean fluorescent intensity, in green calibrated units [GCU]) were analysed with the basic analyser module of the IncuCyte Zoom software. Graphed values are based on the mean values in each well, and standard error of measurement was calculated using the mean values from the triplicate wells for each replicon.

Supplementary Fig. S1 Green fluorescence as a read-out for replication of tagged replicons. Replication of tagged replicons in the background of CHIKV repl sg-ZsGreen in
HuH-7 cells. After seeding HuH-7 cells into 12-well plates and an overnight incubation period to allow cells to adhere, we used lipofection to transfect CHIKV replicon RNA.
We compared replication of CHIKV SNAP-P3 sg-ZsGreen (SNAP) and CHIKV mCherry-P3 sg-ZsGreen (mCherry) to that of CHIKV repl sg-ZsGreen (WT), which encodes untagged, wild-type nsP3. and low-intensity structures (pink) are displayed sequentially. To illustrate differences in fluorescence intensity of nsP3 clusters, we increased the contrast sequentially in the volume rendering mode before overlaying the 3-D surfaces of low-intensity clusters (pink). The video incorporates zoomed-in views from Fig. 7b, focusing on "Zoom 1" and "Zoom 2" regions.