key: cord-0963950-0p8485g8
authors: Milewska, Aleksandra; Owczarek, Katarzyna; Szczepanski, Artur; Pyrc, Krzysztof
title: Visualizing Coronavirus Entry into Cells
date: 2020-05-11
journal: Coronaviruses
DOI: 10.1007/978-1-0716-0900-2_18
sha: bd5b16d1afd0f6e1215acd3b0d65d03beb553b85
doc_id: 963950
cord_uid: 0p8485g8

Coronavirus entry encompasses the initial steps of infection, from virion attachment to genome release. Advances in fluorescent labeling of viral and cellular components and confocal imaging enable broad spectrum studies on this process. Here, we describe methods for visualization of coronavirus entry into immortalized cell lines and 3D tissue culture models.

Coronavirus entry is initiated by interaction between the trimeric spike (S) protein and its receptor, which is expressed on the surface of the susceptible cell. During entry, the S protein undergoes structural rearrangement, which brings the cellular and viral membranes into proximity to mediate fusion. Such a structural switch may be triggered by different stimuli, including receptor binding, proteolytic cleavage of the S protein, and/or acidification of the microenvironment. The requirement for specific stimuli is speciesdependent, and consequently, different coronaviruses enter cells at various subcellular sites. Some coronaviruses fuse at the plasma membrane, whereas others are believed to enter the cell through receptor-mediated endocytosis, followed by fusion within the endosomal compartments. Furthermore, recent reports show that the entry portal may vary depending on the tissue and the cell. These differences may affect the host range, pathogenicity, and cell/tissue specificity [1] .

To visualize coronavirus entry into cells, we developed a confocal microscopy-based analysis method and single-virus particle-tracking tools. We believe that coronavirus entry is highly dependent on the in vitro model used, and therefore, we study coronavirus entry using susceptible cell lines, but we also confirm our observations using the complex ex vivo model of human airway epithelium (HAE). HAE cultures are formed by the multilayered, fully differentiated primary human airway epithelial cells grown at the air/liquid interface built on collagen-coated plastic supports. HAE cultures mimick the natural conductive airway epithelium and serve as the most reliable model to study coronavirus infection [2] [3] [4] [5] [6] .

To study coronavirus entry, we first prepared concentrated stocks of coronaviruses. We found the iodixanol-based medium to be optimal for coronavirus concentration and purification, as in other media (e.g., sucrose gradient), virions rapidly lose infectivity ( Fig. 1 ). Single-step iodixanol purification was sufficient for the successful visualization of single-virus particles in the cell (Fig. 2) . We visualized single coronaviral particles using immunodetection, and we decided that staining specifically for the N protein was superior due to the high availability of antibodies and abundance of the protein itself [7, 8] .

In our research, we tracked virus entry from virus attachment to the cell. First, using a variety of techniques, we studied the attachment itself and made an effort to identify attachment receptors, which are usually broadly distributed molecules, e.g., sugar moieties [8] [9] [10] [11] . Using confocal microscopy-based analysis complemented with flow cytometry appeared to be the optimal means to investigate these processes [11] [12] [13] . Further, we aimed to identify entry receptors and virus-receptor interactions. Except for proteomic analyses, which are not covered in this chapter, confocal analysis of virus-receptor interaction allowed us to delineate the process of coronavirus entry [14, 15] . Next, we checked whether the virus undergoes fusion on the surface of the cell or first requires endocytic internalization. For this we employed chemical inhibitors hampering alteration of the endosomal microenvironment (e.g., ammonium chloride, bafilomycin A), which should affect endocytic entry, but not fusion on the cell surface. Further, we tested whether the virus colocalizes with the early endosome antigen 1 (EEA-1) early after internalization. EEA-1 is a hydrophilic protein, which in cells is found exclusively in early endosomes [16] (see Fig. 3 for examples) [17, 18] .

Subsequently, we found that identification of the endocytosis route for a given coronavirus should first rely on determining the colocalization of virions with proteins specific for endosomes HCoV-OC43 strain 0500 on human airway epithelium. Two exemplary pictures are presented for each virus. Viral nucleocapsids are shown in green (Alexa Fluor 488), actin cytoskeleton in red (Atto 633), and nuclei in blue (DAPI). Each image is a single confocal plane. Scale bar 10 μm entering by a given pathway. For this, we incubated permissive cells with a purified coronavirus stock at 4 C to enable virus adhesion to the cell surface and block its internalization, as at this temperature intracellular transport is inhibited. Consequently, we increased virus density on the cell surface, and by increasing the temperature, we were able to synchronize virus entry. Cultures were incubated at higher temperatures and fixed at specific time points. Samples were immunostained for coronaviral and host proteins. The colocalization rate between viral and cellular proteins was estimated using Pearson's and Manders' correlation coefficients. As controls, we used protein cargos, such as transferrin, which is endocytosed via a clathrin-mediated pathway or cholera toxin or albumin, which enter cells via caveosomes [19, 20] (Fig. 4) . We verified virion colocalization with markers for specific entry routes in cells pretreated with chemical inhibitors or cells with silenced expression of proteins required for specific endocytic pathways. In these experiments, endocytosis of the virus or reference cargo was blocked, and internalization was analyzed with confocal imaging. For example, for all coronaviruses studied, we observed efficient inhibition of internalization using dynamin-2 inhibitors. On the other hand, chemicals blocking clathrin-mediated endocytosis affected only human coronavirus (HCoV)-NL63 entry, and caveolin inhibitors hampered canine respiratory coronavirus (CRCoV) and HCoV-OC43 internalization. Furthermore, as entry via endocytosis usually requires re-arrangement of the cytoskeleton, we also tested virus internalization in the presence of inhibitors affecting actin filaments and microtubules [11, 12, [21] [22] [23] [24] [25] [26] [27] (Fig. 5 ). Confocal imaging of cells allows us to visualize not only the process but also to quantify it. The efficiency of entry can be estimated based on the ratio between internalized viruses and viruses attached to the surface, e.g., in the presence and absence of an inhibitor.

Virus internalization is not always linked with infectious entry. As we have shown in Owczarek et al [22] , the re-direction of the virus into the micropinocytosis pathway allows effective entry but does not allow for fusion and infectious entry. For that reason, it is of importance to verify whether effective virus entry relies on specific pathways. We therefore always complement experiments described above with studies on virus replication in the presence or absence of chemical inhibitors of specific pathways [21, 22] . Nystatin Sterol-binding agent 10-100 μg/ml 50 μg/ml 5 μg/ml [49] ND no data. Working concentration optimized for the stated cell lines in our laboratory 7. 0.2 U/ml Phalloidin labeled with Alexa Fluor 633.

8. Antifade medium (e.g., ProLong Gold antifade medium).

9. Immersion oil (e.g., Immersol 518 F), 10 . Rotary shaker.

11. Fine tipped forceps.

12. Scalpel.

13. Glass slides.

14. Nail polish. 2. Image analysis software capable of colocalization analysis and counting three dimensional objects, e.g., ImageJ1.52i [50, 51] .

All procedures involving infectious material should be performed inside a biological safety cabinet until the fixing step. 9. Cut out the membrane from the insert with a scalpel or push out using a tube. Mount on a glass slide with the antifade medium. Cover the membrane with a coverslip and seal with nail polish. Let it dry overnight at RT.

1. Check the confocal microscope to determine which light sources are available (emission wavelengths).

2. Choose fluorophores. Fluorescent dyes not only differ in the maximal excitation and emission wavelengths but also exhibit different spectrum width and quantum yield (which is a measure of the efficiency of photon emission compared to the number of photons absorbed). We have successfully visualized up to four fluorophores at the same time (cells' nuclei stained with DAPI + coronaviral proteins stained with Alexa Fluor 488 + host entry molecules stained with Alexa Fluor 546 + actin cytoskeleton stained with Atto 647 or Atto 633), but a higher number of fluorophores is possible.

To analyze the excitation and emission spectra and model compensation matrixes for evaluation of the spectral compatibility of fluorescent dyes and probes, online tools, e.g., Fluorescence SpectraViewer (www.lifetechnologies.com/handbook/ spectraviewer) may be used. 2. Choose the scanning mode. Simultaneous scanning of different wavelengths in samples labeled with multiple fluorophores is faster, but often falsifies the signal (cross talk between channels). Therefore, sequential scanning with a single laser line and a single detector activated (single fluorophore) at a time is the method of choice. The region of interest may be scanned by switching tracks line-by-line or frame-by-frame. In the first case, lasers and detectors are changed for each line, so the multicolor image is acquired at once. In the latter case, a single color is captured for the entire frame, and frames representing different fluorophores are superimposed. Scanning line-by-line while faster and favored allows modifying only channel and laser intensities between different fluorophores. In turn, frame-by-frame scanning allows also for changing beam splitters, gain, offset, and pinhole setting between the acquisition of single-color tracks.

3. Set the optical and digital zoom. To ensure adequate spatial sampling, the pixel size should be half of the microscope resolution to ensure Nyquist sampling conditions for all measurements.

4. Adjust the speed of scanning. If the sample is not prone to photobleaching, decreasing the speed will yield higher image quality; however, this may also increase the noise signal, so this step should be optimized. Otherwise, averaging may be applied while the speed of scanning remains high-such settings should improve the signal-to-noise ratio. The low scanning speed and averaging will impact the acquisition time.

5. Set image resolution at 16 bit (65,536 levels of gray, optimal for quantitative measurements).

6. Set laser power, gain, and offset values. To ensure that the signal is within the dynamic range of the detector, live mode along with range indicator should be used. It will also show under-and overexposed pixels. If there is a need to increase the fluorescence signal, it is better to modify the power of the laser or the detector's gain than to apply the digital gain boost. This step demands optimization because too high laser intensity will photobleach the sample, while too high gain will generate noisy images (see Note 13).

To acquire a 3D image of the region of interest, activate the z-stack mode. 2. Adjust the step size. It is appropriate to set a value that fits the Nyquist sampling criteria; however, for some purposes, oversampling may be recommended, e.g., it may help to separate high-frequency noise from the signal during image deconvolution. Our standard setting is 150 nm because it helps us to distinguish viruses by increasing signal to noise ratio in the z-axis.

A number of image analysis software packages are available. In our research, we use the free software package ImageJ.

It is easy to misrepresent data when adjusting the images. Therefore, it is essential to understand the difference between acceptable and unacceptable image adjustment. This is especially important for confocal microscopy or superresolution microscopy, where the presented image is always an interpretation of the signal imposed by the operator.

1. Limit adjustments as much as possible. Linear adjustments of brightness and contrast are generally considered safe and do not require an explanation as long as they are applied uniformly to all pictures in the set. Changes that do not affect each pixel in the same way (e.g., gamma) or more radical changes while allowed should be clearly described and justified in the Methods section and mentioned in the figure caption (see Note 15).

2. Even basic manipulations, as contrast adjustments, may remove some or introduce new elements into image, falsifying the message. When selecting parameters, always refer to mocktreated cells and suitable isotype controls.

3. The reader assumes that the picture presented in a figure represents a single microscope field. Therefore, always make clear divisions between elements from different images. A comparison of different images or their fragments may require resizing to obtain matching resolutions across the complete figure.

While downsizing requires averaging of adjacent pixels and is acceptable, upsizing of the image introduces artificial pixels to the image and should be avoided.

4. Often, when imaging single virus particles, the number of events is low in a single field of view. To cope with this issue, imageJ Z functions allow for a combination of multiple slices into one using one of six different projection methods.

It is not possible to determine whether the virus entered the cell or is retained on the surface using only xy images, but it is necessary to analyze orthogonal views. The reslice option allows for the generation of xz and yz planes and also allows for downstream analyses and adjustments.

6. Never mix different color models. Doing so splits the signal into two channels, adding some of it to pre-existing channels (e.g., yellow added to RGB is split to red and green channels) and may yield false colocalization.

1. The ImageJ JACoP plugin is used for the calculation of colocalization. To use it, images should be split into single channels and loaded in pairs to be compared.

2. Select the method. There are a number of parameters that describe the colocalization of objects, and we find Pearson's and Mander's coefficients most convenient. While Pearson's correlation is easy and fast, Mander's coefficient allows to determine how object A colocalizes with object B, and how object B colocalizes with object A. To illustrate the importance of such distinction, one may imagine that while all virions (A) colocalize with clathrin (B) while entering the cell, not all clathrin-coated vesicles (B) carry a virus (A), and some clathrin is dispersed in the cell.

3. When calculating Mander's coefficient, it is essential to adjust the threshold for both channels-in contrast to Pearson's, it is necessary to define background signal before analysis (see Note 16).

1. Split the image into separate channels.

2. Calculate the total number of viruses using, e.g., 3D Objects Counter tool (Fiji ImageJ)-adjust the threshold value using mock-treated cells and remove noise and aggregates using minimum and maximum size filters.

3. Calculate the total number of cells using the DAPI channel and the 3D Objects Counter tool-adjust threshold value and minimum size filter to remove the signal from cell debris. As this analysis is prone to error while analyzing pictures of dense cultures (e.g., HAEs), select for the option to create a surface map, and revise obtained results.

4. Divide the number of virus particles by the number of cells to obtain an average number of particles per cell.

To calculate the number of particles in the nuclei of the cell, one may create a mask using the DAPI channel and subtract parts of other channels using an image calculator tool before counting. Unfortunately, masks created using the signal for actin are not sufficient to distinguish internalized and surface viral particles and they should be counted using other methods (e.g., manually). To reduce uncertainty, use gamma on the actin channel to obtain surface without gaps.

1. PFA preparation procedure, purity, and supplier may affect the efficacy and specificity of the immunostaining. This step requires optimization.

2. Cytotoxic activity of chemical inhibitors may affect coronavirus entry into cells in a nonspecific way. A cell viability assay for each compound should be carried out in the experimental setting before the actual experiment.

3. Obtain HCoV-NL63 stock by infecting monolayers of LLC-Mk2 cells in T75 flasks and maintaining at 32 C. Obtain a HCoV-OC43 stock by infecting monolayers of HCT-8 cells in T75 flasks and maintaining at 32 C. Obtain a stock of CRCoV by infecting monolayers of HRT-18G cells in T75 flasks and maintaining at 37 C. To asses virus yield, titrate on fully confluent permissive cells, according to the method of Reed and Muench [52] . For CRCoV, additional staining is required as the virus does not produce cytopathic effect on HRT-18G cells [12] . 4 . To obtain HCoV stocks, lyse cells by two freeze-thaw cycles and aliquot at À80 C. To obtain CRCoV stock, remove the supernatant, scrape the cells, and freeze-thaw them in a low volume of liquid. Mix with previously removed supernatant and aliquot at À80 C.

5. For some coronaviral strains (e.g., clinical isolates), the stock titer may be low and insufficient for immunostaining. In such cases, virus stock should not be frozen but processed immediately after collection. It is also possible to scale up the culture and concentrate the media to 8 ml.

6. We observed that some inhibitors are active in some cells while showing no activity in others. It is essential to include reference compounds in each study.

7. The fusogenic activity of some viruses can be inhibited by acidic buffer (0.1 M NaCl, 0.1 M glycine, pH ¼ 3; "acid wash"). This may be useful if one is willing to test which virions were on the cell surface at the time of exposition. For this, cells are washed three times with cold acidic buffer, followed by washing with PBS (pH ¼ 7.4).

8. Importantly, coronavirus-infected samples will be treated as non-infectious after incubation in PFA. For that reason, it is important to transfer the cells on coverslips to new non-contaminated plates for fixation. For ease, fixative can be added to wells prior to the transfer of coverslips.

9. The optimal time for each route to be tested is best optimized for each cell type by testing various time points. However, a good starting point is a timecourse study, taking samples every 20-30 min from 0 to 180 min, perhaps longer depending on the endocytic marker being used. Early events are often best studied in 5-min increments. If seeking to understand whether the virus is colocalising with a marker of an endosome (e.g., EEA1), colocalisation might be expected a bit later in this timecourse.

10. Sequential immunostaining of different epitopes while lengthening the procedure allows for optimization of conditions for each antibody and allows for some atypical staining combinations (e.g., first staining with rabbit antibody specific to protein X and a secondary goat anti-rabbit antibody followed by blocking and second staining with mouse antibody specific to protein Y and secondary rabbit anti-mouse antibody).

11. Alternatively, insert membranes may be cut out after fixation, cut into 2-4 pieces and perform subsequent steps to stain for different antigens in microcentrifuge tubes.

12. Stabilize the temperature of the sample and the microscope's environment before image acquisition. All materials should be left to equilibrate the temperature for at least 15 min. This will help to avoid image deformations caused by temperature drift.

13. If slides are imaged during different sessions, acquisition settings should be identical (tip: they are usually saved as metadata in the image file). It is also essential to verify the values set by manual knobs, e.g., laser power regulation. Often they function independently of the software and will not be adjusted automatically.

14. When counting particles from cultures propagated on Thin-Cert insert membranes, exclude few bottom slices as high membrane autofluorescence tends to impair 3D Objects Counter algorithm function.

15. ImageJ scale bars during generation delete the fragment of the picture they cover. This process is irreversible, so it is good practice to store original versions of photos for future use. 16 . The antibody size is in a range of 10-15 nm, which may severely affect the results, especially in superresolution microscopy; to improve the localization analysis, smaller labels, such as aptamers, nanobodies, or quantum-dots, may be used.

Cell lines: LLC-Mk2 (ATCC: CCL-7), HCT-8 (ATCC: CCL-244), HRT-18G (ATCC: CRL-11663)

Antibiotics (1Â): penicillin (100 U/ml), streptomycin (100 μg/ml), ciprofloxacin (5 μg/ml)

DMEM: Dulbecco's modified Eagle's medium, supplemented with 3% heat-inactivated fetal bovine serum (FBS) and antibiotics (1Â)

MEM: two parts Hanks' MEM, one part Earle's MEM, supplemented with 3% heat-inactivated fetal bovine serum (FBS), and antibiotics (1Â)

RPMI: RPMI-1640, supplemented with 3% heat-inactivated fetal bovine serum (FBS) and antibiotics (1Â)

Cell culture incubator set at 32 or 37 C and with 5% CO 2 . 8. Virus stock (HCoV-NL63, HCoV-OC43 or CRCoV)

Centrifugal filters (10,000 kDa cut off)

Phosphate-buffered saline (PBS)

Iodixanol solution (Optiprep medium)

Ultracentrifuge vials

Ultracentrifuge capable of reaching 170,000 Â g

Denaturing sample buffer for SDS-PAGE. 1. Washing buffer: 0.5% Tween 20 in PBS. 2. Blocking buffer: 5% BSA in PBS. 3. Dilution buffer: 1% BSA and 0.5% Tween 20 in PBS. 4. Primary and secondary antibodies

Fields virology, 6th edn

Efficient replication of the novel human betacoronavirus EMC on primary human epithelium highlights its zoonotic potential

Culturing the unculturable: human coronavirus HKU1 infects, replicates, and produces progeny virions in human ciliated airway epithelial cell cultures

Well-differentiated human airway epithelial cell cultures

Infection with human coronavirus NL63 enhances streptococcal adherence to epithelial cells

Use of sensitive, broad-spectrum molecular assays and human airway epithelium cultures for detection of respiratory pathogens

The nucleocapsid protein of human coronavirus NL63

Novel coronavirus-like particles targeting cells lining the respiratory tract

Membrane protein of HCoV-NL63 is responsible for interaction with the adhesion receptor

Human coronavirus HKU1 spike protein uses O-acetylated sialic acid as an attachment receptor determinant and employs hemagglutininesterase protein as a receptor-destroying enzyme

Human coronavirus NL63 utilizes heparan sulfate proteoglycans for attachment to target cells

Canine respiratory coronavirus employs caveolin-1-mediated pathway for internalization to HRT-18G cells

Canine respiratory coronavirus, bovine coronavirus, and human coronavirus OC43: receptors and attachment factors

HTCC: broad range inhibitor of coronavirus entry

Novel polymeric inhibitors of HCoV-NL63

EEA1, an early endosome-associated protein. EEA1 is a conserved alpha-helical peripheral membrane protein flanked by cysteine "fingers" and contains a calmodulinbinding IQ motif

Rab7 is functionally required for selective cargo sorting at the early endosome

Mobility of late endosomal and lysosomal markers on phagosomes analyzed by fluorescence recovery after photobleaching

Saturation of the endocytic pathway for the transferrin receptor does not affect the endocytosis of the epidermal growth factor receptor

Non-coated membrane invaginations are involved in binding and internalization of cholera and tetanus toxins

Entry of human coronavirus NL63 into the cell

Early events during human coronavirus OC43 entry to the cell

Identification of a new human coronavirus

Human coronavirus NL63 employs the severe acute respiratory syndrome coronavirus receptor for cellular entry

Interaction between the spike protein of human coronavirus NL63 and its cellular receptor ACE2

Attachment factor and receptor engagement of SARS coronavirus and human coronavirus NL63

The novel human coronaviruses NL63 and HKU1

Bafilomycin inhibits proton flow through the H+ channel of vacuolar proton pumps

The bafilomycin/concanamycin binding site in subunit c of the V-ATPases from Neurospora crassa and Saccharomyces cerevisiae

An overview of inhibitors of Na(+)/H(+) exchanger

Amiloride inhibits macropinocytosis by lowering submembranous pH and preventing Rac1 and Cdc42 signaling

Wortmannin is a potent phosphatidylinositol 3-kinase inhibitor: the role of phosphatidylinositol 3,4,5-trisphosphate in neutrophil responses

Dynasore, a cell-permeable inhibitor of dynamin

Iminochromene inhibitors of dynamins I and II GTPase activity and endocytosis

Dynamin2, clathrin, and lipid rafts mediate endocytosis of the apical Na/K/2Cl cotransporter NKCC2 in thick ascending limbs

Tetraspanin CD9 regulates cell contraction and actin arrangement via RhoA in human vascular smooth muscle cells

Cytochalasin D inhibits actin polymerization and induces depolymerization of actin filaments formed during platelet shape change

An isoform-selective, small-molecule inhibitor targets the autoregulatory mechanism of p21-activated kinase

New anti-actin drugs in the study of the organization and function of the actin cytoskeleton

NSC23766, a widely used inhibitor of Rac1 activation, additionally acts as a competitive antagonist at muscarinic acetylcholine receptors

Effect of ROCK inhibitor Y-27632 on normal and variant human embryonic stem cells (hESCs) in vitro: its benefits in hESC expansion

Pharmacological properties of Y-27632, a specific inhibitor of rho-associated kinases

Nocodazole action on tubulin assembly, axonal ultrastructure and fast axoplasmic transport

Stabilization of clathrin coated vesicles by amantadine, tromantadine and other hydrophobic amines

Mis-assembly of clathrin lattices on endosomes reveals a regulatory switch for coated pit formation

Pitstop 2 is a potent inhibitor of clathrin-independent endocytosis

Filipindependent inhibition of cholera toxin: evidence for toxin internalization and activation through caveolae-like domains

Cholesterol depletion using methyl-β-cyclodextrin

How do the polyene macrolide antibiotics affect the cellular membrane properties?

Fiji: an open-source platform for biological-image analysis

A guided tour into subcellular colocalization analysis in light microscopy

A simple method of estimating fifty per cent endpoints