Super resolution microscopy and deep learning identify Zika virus reorganization of the endoplasmic reticulum

Abstract

The endoplasmic reticulum (ER) is a complex subcellular organelle composed of diverse structures such as tubules, sheets and tubular matrices. Flaviviruses such as Zika virus (ZIKV) induce reorganization of ER membranes to facilitate viral replication. Here, using 3D super resolution microscopy, ZIKV infection is shown to induce the formation of dense tubular matrices associated with viral replication in the central ER. Viral non-structural proteins NS4B and NS2B associate with replication complexes within the ZIKV-induced tubular matrix and exhibit distinct ER distributions outside this central ER region. Deep neural networks trained to distinguish ZIKV-infected versus mock-infected cells successfully identified ZIKV-induced central ER tubular matrices as a determinant of viral infection. Super resolution microscopy and deep learning are therefore able to identify and localize morphological features of the ER and allow for better understanding of how ER morphology changes due to viral infection.

Towards a Quantitative Single Particle Characterization by Super Resolution Microscopy: From Virus Structures to Antivirals Design

“…Furthermore, from the cell interaction to the assembly, proteins could be monitored and quantified, identifying the clusters of proteins and their role in the infection of viruses. Additionally, SRM could be implemented in the relevant study of the molecular changes triggered by viral infection inside cells, such as the reorganization of the endoplasmic reticulum given by Zika infection ( Long et al, 2020 ), the modification of the membrane due to the filament formation of influenza ( Kolpe et al, 2019 ) or the spatial re-arrangement of ESCRT machinery by HIV-1 ( Bleck et al, 2014 ).…”

A novel highly potent inhibitor of TMPRSS2-like proteases blocks SARS-CoV-2 variants of concern and is broadly protective against infection and mortality in mice

“…Camostat mesylate was obtained from MilliporeSigma. The SARS-CoV-2 nucleocapsid antibody [HL344] (GTX635679) was kindly provided by Genetex; mouse anti-dsRNA antibody (J2-1904) was purchased from Scions English and Scientific Consulting 34 ; Hoechst 33258 and secondary antibodies goat anti-mouse IgG Alexa Fluor 488 (A11001) and goat anti-rabbit IgG Alexa Fluor 555 (A21428) were obtained from Invitrogen.…”

“…Calu-3 cells were pretreated with 100 nM of the compounds for three hr prior to infection. Cells were fixed and immunofluorescently stained for dsRNA, a marker of viral replication 34 , and for the viral nucleocapsid, a marker of viral entry and translation 35 (Figure S3). Fluorescent highcontent imaging and relative quantification of virally infected cells demonstrated consistent inhibitory profiles across dsRNA and nucleocapsid staining, which mirrored the inhibitory profile observed in the TMPRSS2 proteolytic activity assay (Figure 2A versus Figure 1B).…”

Challenges facing quantitative large-scale optical super-resolution, and some simple solutions

“…It is therefore conceivable that the high level of information contained in SRM images could be exploited in a similar manner. Such applications are beginning to be realized with conventional fluorescence microscopy and only very recently with SRM ( Kraus et al., 2017 ; Laine et al., 2018 ; Long et al., 2020 ; Lu et al., 2018 ). In one such study, SIM combined with deep learning to image and classify a large population of viruses.…”

“…By measuring the morphological features of the various classification groups, new connections could be drawn between viral structures and their mechanism of action ( Laine et al., 2018 ). In another study, a neural network was trained to differentiate 3D STED images of healthy versus Zika-virus-infected cells, and identified morphological changes to the ER associated with viral infection ( Long et al., 2020 ). Although promising, a major limitation of deep learning applications is that the training of a neural network requires large datasets of high-quality images, which are not always readily available.…”