Ultrastructural analysis of SARS-CoV-2 interactions with the host cell via high resolution scanning electron microscopy

SARS-CoV-2 is the cause of the ongoing COVID-19 pandemic. Here, we investigated the interaction of this new coronavirus with Vero cells using high resolution scanning electron microscopy. Surface morphology, the interior of infected cells and the distribution of viral particles in both environments were observed 2 and 48 h after infection. We showed areas of viral processing, details of vacuole contents, and viral interactions with the cell surface. Intercellular connections were also approached, and viral particles were adhered to these extensions suggesting direct cell-to-cell transmission of SARS-CoV-2.


Scientific RepoRtS
| (2020) 10:16099 | https://doi.org/10.1038/s41598-020-73162-5 www.nature.com/scientificreports/ After internalization and RNA release into the cytoplasm, a set of proteins is synthesized triggering the formation of vesicles that become a viral platform ensuring efficient replication and transcription of the RNA 11,12 . New coronavirus particles are assembled in the endoplasmic reticulum and Golgi complex. Membrane budding between these compartments was reported in association with N protein and genomic RNA along with M, E, and S proteins. The complete virions are delivered to the extracellular environment following a conventional secretory route [13][14][15] .
The research community has sought to better understand the genetic makeup of the virus and thus discover how to effectively treat it. Social isolation for 14 days is the main way to prevent the disease from spreading. Quarantine and lockdowns were implemented in cities with high rates of infection and mortality 3 . Death is common in patients with severe symptoms including shortness of breathing, fever, lethargy, respiratory failure, and/or thrombosis 16,17 .
Understanding the virus-cell interactions is key to vaccines, treatments, and diagnoses. Most microscopic studies of SARS-CoV-2 were performed with transmission electron microscopy. Here, we used high resolution scanning electron microscopy (SEM) to study inner cellular structures. The results offer evidence of infectioninduced cellular remodeling and the formation of a specialized region for viral morphogenesis. We also show intercellular extensions for viral cell surfing. These observations offer new insights into the transmission of SARS-CoV-2.

Material and methods
Cells and virus. SARS-CoV-2 isolate (HIAE-02: SARS-CoV-2/SP02/human/2020/BRA (GenBank accession number MT126808.1) was used in this work. The virus was grown in Vero cells (Monkey African Green kidney cell line -ATCC CCL-81) in the Laboratory of Molecular Virology, at Federal University of Rio de Janeiro, Brazil. Vero cells were maintained in DMEM supplemented with 5% fetal bovine serum (FBS; GIBCO) at 37 °C and 5% CO 2 . All work involving infectious SARS-CoV2 was performed in a biosafety level (BSL)-3 containment laboratory.
Infection assays. Semi-confluent (70%) cells were grown on sterile glass coverslips in 24-well tissue culture plates infected with MOI (multiplicity of infection -the rate of virus per cell) values of 0.01, 0.1, or 1 using SARS-CoV-2 in free-serum medium. Fresh medium containing 5% FBS was added after an absorption period of 1.5 h at 37 °C and 5% CO 2 . Cells were processed for electron microscopy 2 or 48 h post-infection (hpi).
High resolution scanning electron microscopy. After 2 or 48 h post-infection (hpi), samples were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2) for 2 h. The coverslips were washed with 0.1 M sodium cacodylate buffer and post-fixed for 40 min in 1% OsO 4 with 0.8% potassium ferrocyanide. After another washing cycle, the samples were dehydrated through a series of increasing concentration (30%-100%) of ethanol. The samples were critical-point-dried in liquid CO 2 in a BALZERS CPD apparatus before monolayer scraping with a conventional adhesive SCOTCH tape as in previous studies 18 . This technique does not totally remove the plasma membrane, as happens when detergent extraction is used, but provides the exposing of large areas of the inner portion of the cells. They were then sputtered with a 5-nm thick platinum coat in a BALZERS apparatus. Samples were observed using an Auriga ZEISS microscope operated between 1.0 and 1.8 kV.

Results
To identify alterations on the surface of SARS-CoV-2-infected cells, we compared their morphology and the occurrence of surface projections (SP). While we did not detect any significant alteration in cell shape, the presence of SP increased on the surface of infected cells at 2 hpi ( Fig. 1A-C). However, no viral particles were observed adhering to the cell surface or beneath these projections ( Fig. 1D). At 48 hpi, we compared the surfaces of mock and infected cells (MOI of 0.1) to highlight the presence of viral particles adhered to the smooth cell surface and to the SP (Fig. 1E, F).
At the same time, and with a MOI of 1, viruses that egressed from a previous cycle of infection were observed during the process of attachment to the cell plasma membrane ( Fig. 2A). The corona-like features of the SARS-CoV-2 particles were discernible via SEM (Fig. 2B), and the measurements showed sizes of approximately 80 nm in diameter (Fig. 2C).
Removal of the host cell plasma membrane before platinum sputtering exposed the interior of the mock and infected cells. While mock cells displayed a diffuse distribution of organelles (Fig. 2D), infected cells exhibited a more polarized disposition of organelles and pit-coated vesicles approximately 100 nm in diameter (Fig. 2E, F). With a MOI of 1, cells at 48 hpi showed a plethora of vacuoles (0.4 to 1 µm; Fig. 3A). These were translocated to the cell plasma membrane presumably to perform exocytosis of viral particles (Fig. 3B). Some of these vacuoles had their content revealed and were filled with immature viruses, amorphous materials, or a hemocyte-like format ( Fig. 3C-E). Although no virus-like particles could be distinguished in the ER, bordering vesicles were observed on the vacuoles (Fig. 3D).
Cells at 48 hpi also had viral particles near the cell surface membrane ruffles (Fig. 4A) and a filopodium-like structure (Fig. 4B). Other viral particles were wrapped with thin (≃ 70 nm) cellular projections that resemble nanotubes (Fig. 4C)

Discussion
Part of the challenge in controlling COVID-19 is the innovative features of this coronavirus. New knowledge on virus genetics and morphology needs to be analyzed concurrently with viral "behavior" within the host cell as well as the dynamics that determine the fate of the particle. To approach SARS-CoV-2/cell interactions, we investigate several steps of virus infection in Vero cells at 2 and 48 hpi by SEM. Vero cells are a widely used model used in viral infection studies and is an adequately supports coronavirus replication 12,14,15,19 . This microscopic approach detailed virus-induced changes in the cell. Our assays were performed using three MOIs (0.01; 0.1 and 1), and we could discern the MOI of 0.1 as the more adequate for this type of study. This MOI allowed the best cell conditions and distribution and also allowed visualization of virions through the cell surface into the cell interior.
The absence of virions adhered to the cells surface at 2 hpi corroborates recent studies performed by Belhaouari et al. 19 in which SARS-CoV-2 particles were only observed at these loci after 12 hpi. In contrast, www.nature.com/scientificreports/ SARS-CoV-2 particles were found lying on the cellular surface at 48 hpi between surface projections and adhered to them. We also observed probable viral particles inside vacuoles suggesting a secretion route. These aggregates of cell organelles and components (Fig. 2F) may reflect the polarized release of virus previously described for SARS-CoV 20 . All viruses measured by SEM display a spiky round shape with a size of around 70-85 nm in diameter considering a platinum coating of 5 nm. This agrees with the dimensions described in recent studies 1,21,22 . Viral particles adhered to smooth surface and microvilli-like surface projetions. The effects on the surface morphology of infected cells varies among viruses. Infection by several viruses including HTLV-IIIB leads to a loss of cell SP that are then replaced by blebs 23 . Microvilli induction or increases were reported in several cases of DNA or RNA viral infection 24,25 . For RNA viruses that egress by budding, e.g., influenza, the increase in SP of infected cells coincides with higher budding rates 26 .
Similar to prior studies on SARS-CoV infection of Vero cells 27 , we also observed a ruffled host cell and thickened edges displaying a layered shape. These sites were appropriate to register the attachment of SARS-CoV-2 particles ( Fig. 2A) similar to transmission electron microscopy images of the same early step of SARS-CoV infection of Vero cells 28 .
Likewise, the proliferation of SP on the infected cells, especially at the apical region of these cells, is similar to SARS-CoV and SARS-CoV-2. In addition, the abundance of SARS-CoV-2 particles held on SP, was recently showed 29 and may facilitate the speed of viral propagation in the epithelium of conducting airways from the lumen of the respiratory superior tract because this environment is colonized by ciliated cells. www.nature.com/scientificreports/ Vacuoles containing viral particles. Cell scraping is a very useful expedient that is occasionally used in studies of host cell/parasite interactions 30,31 . Infected cells are artificially devoid of plasma membranes and exposed to a myriad of vacuoles (Fig. 3A). Drastic vacuolization due to viral infection was previously described for other RNA viruses including SARS-CoV 20,32 . Similar sites were recently reported as virus morphogenesis matrix vesicae (VMMV) 19 . The particles observed in the interior of these VMMVs (Fig. 3C-E) were previously described as doughnut-like particles when observed by electron microscopy 19,33 . SARS-CoV immature particles are presumed to bud into vesicles as part of the assembly process, and thus the observed particles were probably immature viruses devoid of the representative (corona) spikes of this virion. Bordering vesicles were found in close association with the vacuoles (Fig. 3D), and thus we speculate that their role in viral pre-components leads to discharge into the compartments. Studies with other coronaviruses identified large virion-containing vacuoles (LVCVs) where the complete particle would bud. There is correlation between these structures as observed by transmission electron microscopy and our data suggesting the occurrence of both phenomena.

Translocation of vacuoles towards the plasma membrane. Coronaviruses infection leads to mas-
sive remodeling of cell membranes 34,35 ; the more condensed area depicted in the cytoplasm at 48 hpi (Fig. 2F) may correspond to the main locus of viral morphogenesis. The proposed mechanism for the export of viruses to the extracellular space is via fusion of the transport compartment membrane with the cell plasma membrane 20 .
The size of the vacuoles we observed in the cell periphery was not compatible with the identified clathrincoated pits because the vacuoles measure approximately 1 µm; clathrin-coated pits measure near 200 nm in diameter. The presence of these endocytosis-associated players was recently reported in SARS-CoV-2-infected cells. They are likely receptacles to the nucleocapsid after the incoming virus is uncoated 19 .
Thus, our observations suggest that a boost in vacuoles is restricted only to a specific and more condensed part of the cytoplasm. This suggests translocation to the plasma membrane is required for release the viral particles by a fusion mechanism. Cellular bridges containing viral particles. Viral particles adhered to cell surface protrusions that were shown to connect two cells. This observation suggests viral "cell surfing" previously described by other enveloped viruses such as HIV and human metapneumovirus 36,37 . This mechanism is presumed to allow the in vivo penetration of virus in mucosal surfaces that display microvilli-rich cells.
Actin filaments play a fundamental role in viral extrusion by the cell for both RNA and DNA viruses. Actin offers the strength to discharge the progeny virus particles to the extracellular medium, as occurs to some viruses www.nature.com/scientificreports/ that leave the cell by budding, including Fowlpox and West Nile viruses 38,39 . Other examples include actin comets-these are an efficient form of poxvirus dissemination and cell-to-cell HIV spreading, which involves the direct engagement of GAG proteins and F-actin 40,41 .
Previous studies have shown that the cytoskeleton network plays an important role in the maturation and, possibly, in the replication process of SARS-CoV 27 . Communication between the two cells in Fig. 4C-D suggests the occurrence of a thin (< 0.7 µm) strand of F-actin containing tunneling nanotube (TNT). These intercellular membranous connections may provide the transference of molecular information especially viruses 42 . Similarly, virus cell surfing was shown on SARS-CoV-2 infection, which offers new insights into cell-to-cell propagation and virus transmission.  www.nature.com/scientificreports/