Flavivirus Concentrates Host ER in Main Replication Compartments to Facilitate Replication

Abstract Flavivirus remodels the host endoplasmic reticulum (ER) to generate replication compartments (RCs) as the fundamental structures to accommodate viral replication. Here, a centralized replication mode of flavivirus is reported, i.e., flavivirus concentrates host ER in perinuclear main replication compartments (MRCs) for efficient replication. Superresolution live‐cell imaging demonstrated that flavivirus MRCs formed via a series of events, including multisite ER clustering, growth and merging of ER clusters, directional movement, and convergence in the perinuclear region. The dynamic activities of viral RCs are driven by nonstructural (NS) proteins and are independent of microtubules and actin. Moreover, disrupting MRCs formation by small molecule compounds inhibited flavivirus replication. Overall, the findings reveal unprecedented insight into dynamic ER reorganization by flavivirus and identify a new inhibition strategy.


Virus strains
The Zika virus strain (FSS13025) and the DENV-2 virus strain (DENV2_China_SZ_2015) were acquired from the State Key Laboratory of Pathogen and Biosecurity in Beijing, China.

RR-mNeonGreen plasmid construction
The coding sequence of fluorescence protein mNeonGreen was inserted into pcDNA3.1(+)vector.Then DNA coding N-terminal ER retention signal (MHRRRSRSCREDQKPV) with Kozak translation initial sequence was inserted to the upstream of mNeonGreen sequence.And coding sequence of VAMP2 TM region (KYWWKNLKMMIILGVICAIILIIIIVYFST) was introduced to the downstream of mNeonGreen sequence to drive mNeonGreen targeting ER membrane.

Immunofluorescence
HeLa or Huh-7 cells infected by ZIKV/DENV or transfected by NS protein were fixed with 3% PFA containing 0.1% glutaraldehyde at RT for 10 min.After three times of washing by PBS, cells were permeabilized by 0.2% Triton-X100 diluted in PBS at RT for 10 min, and blocked by 3% BSA for 30 min.Then cells were incubated with indicated primary antibodies at 37 ℃ for 2 h followed by secondary antibodies at 37 ℃ for 1 h.Finally, cells were mounted by ProLong Diamond mountant (Invitrogen).EM HeLa cells were seeded in 35-mm dishes and fixed with 2.5 % glutaraldehyde in PBS followed by post-fixation with 1% osmium containing 1.5% potassium ferrocyanide.
Cells were infiltrated with and embedded in SPON812 resin in situ.After polymerization, embedded monolayer cells were cut to 70nm-thick ultrathin sections by diamond knives.The indicated sections were laid on copper grids and double stained with uranyl acetate and lead citrate.EM sample preparation was performed at the Center of Biomedical Analysis, Tsinghua University.And EM images were captured by electron microscope H-7650.
When capturing images of long-term live cells, we chose the largest cross-section of the nuclei as the focal plane and utilized the auto-focusing strategy of DeltaVisionOMX-UltimateFocus.The focusing parameters are set as follows: the move threshold is 100 nm, the maximum iteration is 1, and auto-focusing is conducted every 6 time points.
For nocodazole experiment, ZIKV-infected cells were treated with nocodazole (1μg/ml) for 30min at 16 hpi, and then fresh phenol red-free DMEM was replaced and cell was imaged by SIM.

MRCs (PSR) analysis
The regions of MRCs and whole cell were measured by drawing a freehand ROI around the MRCs and cells using ImageJ.For control cells (uninfected cells), which had no MRCs, a similar ROI in perinuclear region was drawn and defined as perinuclear similar region (PSR).The area percentage of MRCs (PSR) was calculated as the area of MRCs (PSR) divided by that of whole cell.The percentage of ER in MRCs (PSR) was calculated as the integrated density of ER fluorescence (RR-mNeonGreen or ER marker) in MRCs (PSR) regions divided by the that of whole cell (ImageJ).MRCs or PSR from 50 cells were analyzed.

3D-STED imaging
3D-STED images were acquired by Leica TCS-SP8 STED 3X with HC PL APO CS2 100x/1.4oil objective.Serial Z stack sectioning was done with 488nm excitation laser and 592nm STED laser at 140 nm intervals and imaging settings were as follows: xy pixel size 28nm, pinhole 1AU, speed 400Hz.Z-stack images were processed by Lightning module.Imaging data processing 2D-SIM images were deconvoluted by Sparse deconvolution software (developed by Liangyi Chen lab, Institute of Molecular Medicine, Peking-Tsinghua Center for Life Sciences) and time-lapse images were subjected to CorrectBleach plugin in ImageJ (NIH).For 3D images, the color-coding by depth is projected by ImageJ and 3D construction was processed by Imaris x64 9.7.2 software.ER cluster trackingRCs or ER clusters trajectories were generated by Manual Tracking plugin in ImageJ.Tracking was performed on time-lapse images at 10 min per frame with lifetime at least 60 min.The moving velocity of RCs or ER clusters from 5-12 cells was calculated.ER clusters directional movement analysis ER clusters trajectories were made by Manual Tracking as described above and displayed as Lines.The trajectories image was analyzed by Sholl Analysis plugin in ImageJ.Manually mark the center of nuclei and set the ending radius as 700 pixels.Record the number of trajectories intersecting with circles with different radii.ER cluster area growth analysis ER cluster/RCs area was measured by drawing a freehand ROI around the cluster and then record the area of ROI in ImageJ.The growth rate was calculated as (end area-start area)/time.ER clusters from 5-12 cells were recorded and analyzed.Viral RNA quantificationHeLa cells were infected with ZIKV.Total RNA was extracted by RNAfast200 kit (FASTAGEN).RT-qPCR was performed by FastKing one-step RT-qPCR kit (TIANGEN).Viral dsRNA quantification in MRCs or RCs ER labeled by RR-mNeonGreen was segmented by Fiji.For MRCs region identification, ER channel was filtered by Gaussian Blur with Sigma (Radius) 5.00.Then make binary by manually set threshold and select MRC region manually.For RCs identification, ER channel was filtered by Gaussian Blur with Sigma (Radius) 2.00.Then make binary by manually set threshold and select RCs region manually.The integrated signal intensities of dsRNA in MRCs or RCs regions are calculated by Fiji and the dsRNA intensity per unit area was acquired by dividing integreated dsRNA intensity in each MRCs or RCs by the area.***P < 0.001.ns, not significant.

Figure
Figure S4 Nonstructural proteins of ZIKV induce RCs-like structures.

Figure
Figure S5 RCs movement is independent of microtubules and actin.(A) Response of RR-mNeonGreen stable cells to different concentration of nocodazole.Bright-field and RR-mNeonGreen fluorescent images were taken at 2 hours after the treatment of nocodazole.BF, bright-field.Scale bar, 50 μm.(B) Comparison of area growth rate of ZIKV RCs (with or without nocodazole) and ER clusters induced by NS1-5 or NS1-4B.(C) Comparison of motion velocity of ZIKV RCs (with or without nocodazole) and

Figure
Figure S6 Formation of MRCs is important for viral replication.
(A) Bortezomib (20 nM) treatment impaired MRCs formation and significantly decreased viral dsRNA.Scale bar, 20μm.(B) The percentage of ZIKV-infected cells with different types of RCs upon the treatment of TVB-2640.More than 200 ZIKV-infected cells were calculated according to the type of viral RCs.(C) Area of MRCs were measured and plotted upon the treatment of DMSO or TVB-2640 (1 μM).

(
D) Area of RCs were measured and plotted upon the treatment of DMSO or TVB-2640 (1 μM).(E) TVB-2640 (1 μM) inhibited ZIKV replication.Total RNA was extracted from DMSO-or TVB-2640-treated ZIKV-infected cells and the amount of viral RNA was determined by RT-qPCR.