Human epithelial colorectal adenocarcinoma Caco-2 cells (86010202 obtained directly from Sigma Aldrich, St. Louis, MO) were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal calf serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 1% non-essential amino acids and 2 mM L-glutamine (Life Technologies, Grand Island, NY) at 37°C in 5% CO₂. This was done for 7–10 days to allow the cells to become mature, differentiated and establish polarized epithelial monolayers.

The stock for Hazara virus strain JC280 (GenBank accession number M86624.1) was produced in human adrenal cortex adeno carcinoma SW-13 cells (CCL-105 obtained directly from American Type Culture Collection, Manassas, VA) maintained in Leibovitz's L15 medium (L15) supplemented with 10% heat-inactivated fetal bovine serum, 25 mM HEPES, 100 U/ml penicillin, 100 μg/ml streptomycin (Life Technologies) at 37°C in 5% CO₂.

Caco-2 cell monolayers were starved in serum-free DMEM overnight and then infected with Hazara virus at three multiplicities of infection (MOI) 0.02, 1 and 2 for 1 h at 37°C in 5% CO₂ in DMEM supplemented with 2% heat-inactivated fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin. After 1-h infection, cells were rinsed and maintained for 24 or 48 h post infection (hpi) at 37°C in 5% CO₂ in serum free DMEM supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin. Then, the cells were processed for migration assay, imaging, and proteomics.

Caco-2 cells were seeded in μ-dishes with inserts (Ibidi GmbM, Martinsried, Germany) and cultured for 10–12 days to allow the cells to become 100% confluent, mature, differentiated and polarized epithelial monolayers. Then, cells were serum-starved overnight and either non-infected or infected with Hazara virus at MOI 0.02 and 2.0 for 1 h at 37°C in 5% CO₂ in serum free DMEM supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin. After 1-h infection, the cells were rinsed and maintained for 0, 24, or 48 hpi (Figure 2). At 0, 24 and 48 hpi, the insert was removed to get two cell patches with a 500 ± 50 μm cell-free gap in between allowing the cells to migrate and heal the wound. For each dish, the images of cells migrating into the gap area were taken at indicated time points (Figure 2) using the benchtop microscope JuLI (NanoEnTek Inc., Seoul, South Korea); between imaging, cells were returned to the incubator. Migration activity was calculated by measuring the relative area of the image occupied by cells on each dish at each time point with the Image J software (NIH, Bethesda, MD https://imagej.nih.gov/ij/). At least 3 independent experiments were performed on separate days on different cell passages.

Determination of the concentration of Hazara virus in the virus stock and in the monolayers after migration experiments was performed as described previously (Andersson et al., 2004). Briefly, SW-13, or Caco-2 cells in 96-well plates were infected with 10-fold serial dilutions of the virus stock for 1 h at 37°C in 5% CO₂. Alternatively, Caco-2 monolayers in μ-dishes after closure of the wound were used. Cells were fixed, permeabilized, labeled with polyclonal rabbit anti-CCHFV/Hazara N antibodies (Andersson et al., 2004) and Alexa Fluor 488-conjugated polyclonal goat anti-rabbit antibodies (A11008, Life Technologies). Fluorescent foci were counted in an Axiovert 35 fluorescence microscope (Carl Zeiss, Jena, Germany) equipped with a ProgResC camera (Jenoptik, Jena, Germany). This allowed the determination of the infectious virus titer in focus forming units per ml (FFU/ml).

Caco-2 monolayers grown on glass coverslips (thickness 0.17 ± 0.01, 13 mm-diameter; Karl Hecht Assistent, Sondheim, Germany) were washed with PBS pH 7.3 and pretreated with 0.2% Triton X-100 (Sigma Aldrich) in PBS for 2 min on ice. The samples were then fixed in 2.5% paraformaldehyde (Sigma Aldrich) in PBS pH 7.3 for 20 min at room temperature. The pre-treatment with 0.2% Triton X-100 was important for a clear labeling of proteins. After washing with PBS cells were further permeabilized in 0.05% Triton X-100 in PBS for 5 min and washed again. Non-specific background staining was blocked for 10 min in PBS containing 1% BSA and 10 mM glycine. The washing was repeated, and polyclonal rabbit or monoclonal mouse antibodies against CCHFV/Hazara N (Andersson et al., 2004), monoclonal mouse antibodies against IQGAP1 or polyclonal rabbit antibodies against AQP6 (05-504, AB3073, Millipore, Temecula, CA), diluted 1:200 in blocking buffer were then applied for 1 h at 37°C in a moist chamber. After washing, Alexa 568-conjugated goat anti-rabbit and Alexa 488-conjugated goat anti-mouse antibodies (A11036, A11029, Life Technologies) diluted 1:400 were added and incubated for 1 h at 37°C in a moist dark chamber. To detect F-actin, cells were stained with Alexa 488-conjugated phalloidin (A12379, Life Technologies). Nuclei were labeled with DAPI (Life Technologies), according to the manufacturer's instructions. Finally, coverslips were mounted in ProLong Gold (Life Technologies). The specimens were examined through 63x oil immersion objectives with NA 1.40 in a fluorescence microscope Zeiss Axio Observer Z1 with the confocal system Zeiss LSM700 and Zeiss ZEN software (Carl Zeiss, Jena, Germany). Fluorescence intensity in juxta-membrane regions in the cell monolayers and intensity profile plots across the cell monolayers were measures and quantified using the ImageJ software (NIH). At least 3 independent experiments were done on separate days on different cell passages.

Caco-2 monolayers grown on 6-well plates or flasks were washed with PBS, pH 7.6 and lysed with ice-cold RIPA buffer (150 mM NaCl, 1% deoxycholic acid sodium salt, 1% N-40, 0.1% SDS, 10 mM EDTA pH 8.0, 10 mM Tris pH 7.4 dissolved in PBS) supplemented with 25U nuclease (Thermo Scientific, Rockford, IL), 1 mM phenyl-methyl-sulfonyl-fluoride, 1 mM Na₃VaO₄, 25 mM NaF (Sigma Aldrich), protein inhibitors Complete (Roche Diagnostics, Mannheim, Germany). Cell suspensions were homogenized through a 21-gauge needle and centrifuged at 18,000 g for 30 min at 4°C, and the supernatants were collected. The protein concentration in cell lysates was measured with the Bio-Rad D_C protein assay (Bio-Rad Laboratories, Hercules, CA). The samples were further diluted in Laemmli sample buffer at equal protein concentrations, heated for 5 min at 95°C and then subjected to electrophoresis. They were loaded on 8–16% SDS-polyacrylamide gels (Lonza, Rockland, ME), and after separation, proteins were electrophoretically transferred to a PVDF Immobilon-FL membrane (Millipore); the quality of the transfer was controlled by Ponceau S staining (Sigma Aldrich). Non-specific binding was blocked by 1-h incubation in 5% non-fat milk in PBS, pH 7.6 containing 0.18% Tween 20 at room temperature. The membranes were then incubated with antibodies against CCHFV/Hazara N (Andersson et al., 2004), IQGAP1, AQP6, and GAPDH (05-504, AB3073, MAB374 Millipore) diluted 1:1000 in blocking buffer overnight at 4°C. After washing, they were treated for 1 h at room temperature with IRDye 800CW goat anti-rabbit or IRDye 680CW goat anti-mouse antibodies (926–32,211, 926–68,070, LI-COR Biosciences, Cambridge, UK), diluted 1:10,000 and washed extensively. The signals were detected and the density ratio was quantified using the Odyssey CLx and the Image Studio software (LI-COR). At least 4 independent experiments were performed on separate days on different cell passages.

Total-cell lysates were pre-cleared for 30 min at 4°C with of protein G-Sepharose “4 fast flow” (GE Healthcare, Uppsala, Sweden) and centrifuged before the protein concentration in the supernatants was determined with the Bio-Rad D_C protein assay. Samples with equal protein concentrations were then precipitated overnight at 4°C with 1 μg antibodies against CCHFV/Hazara N (Andersson et al., 2004). Immune complexes were captured at 4°C overnight using protein G-Sepharose “4 fast flow.” The beads were collected by pulse centrifugation and washed three times with cold PBS, pH 7.6. Sepharose beads were re-suspended in Laemmli sample buffer, boiled for 5 min at 95°C, collected by centrifugation, and the supernatant was subjected to 8–16% SDS-PAGE (Lonza). The gels were stained with Coomassie Blue (Thermo Scientific). At least 4 independent experiments were done on separate days on different cell passages.

Coomassie Blue-stained protein bands were excised, reduced, alkylated and digested as described previously (Shevchenko et al., 2006); generated peptides were dried, dissolved in 0.1% (v/v) formic acid in water and analyzed by LC-MS/MS. Peptides were separated by reverse phase chromatography on a 20 mm × 100 μm C18 pre column followed by a 100 mm × 75 μm C18 column with particle size 5 μm (NanoSeparatoons, Nieuwkoop, Netherlands) at a flow rate 300 nL/min. EASY-nLC II (Thermo Scientific) by gradient of 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B) as follows: 0–30% B in 50 min, 30–100% B in 40 min. Automated online analyses were performed with a LTQ Orbitrap Velos Pro hybrid mass spectrometer (Thermo Scientific) with a nano-electrospray source.

Raw files were analyzed using Sequest HT in Proteome Discoverer (Thermo Fisher Scientific, San Jose, CS version 1.4.0.288) against a Uniprot Human database available at UniProtKB website (http://www.uniprot.org/taxonomy/9606) with the following parameters: trypsin as a digestion enzyme; maximum number of missed cleavages 2; fragment ion mass tolerance 0.60 Da; parent ion mass tolerance 10.0 ppm; fixed modification, carbamidomethylation of cysteine; variable modifications, methionine oxidation.

Identified proteins were validated using SCAFFOLD (Version 4.4.8; Proteome Software Inc., Portland, OR). Identifications were based on a minimum of 2 unique peptides, minimum 80% peptide identification probability (using the Scaffold Local FDR algorithm), and minimum 95% protein identification probability (using the Protein Prophet algorithm (Nesvizhskii et al., 2003), resulting in a 0.0% decoy FDR). Proteins that contained similar peptides and which could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. The label-free quantitative analysis of peptides was performed by spectral counting analysis using normalized spectral abundance factor (NSAF) calculated for each protein to normalize run-to-run variations (Zybailov et al., 2006), and quantitative differences were statistically analyzed by two-tailed Student's t-test. Differences with P < 0.1 were considered statistically significant. Identified proteins were categorized according to gene ontology terms.

The interactions between the identified cellular proteins were analyzed by the Search Tool for the Retrieval of Interacting Genes and Proteins, STRING 10.0 (http://www.ncbi.nlm.nih.gov/pubmed/18940858) using medium confidence score 0.4 and all active interaction sources (Franceschini et al., 2013). STRING and NCBI GO annotations by SCAFFOLD analyses were used to group proteins into functional classes.

Data in the graphs are presented as mean ± SE. Statistical analyses are based on two-tailed Student's t-test. The numbers (n) are specified in the figure legends. P < 0.05 (^*), < 0.01 (^**), and < 0.001 (^***) were considered significant. The experiments on migration and imaging were done at least 3 times on separate days on different cell passages. The experiments for immunoblotting, immunoprecipitation, proteome and interactome were repeated 4 times on separate days. Data evaluation and quantification for proteome and interactome experiments are described above.

When viruses target and enter epithelial cells (Connolly-Andersen et al., 2007), a further perturbation of cell properties and wound-healing process may occur. Therefore, we investigated whether an infection with Hazara virus affected epithelial cell migration, using Ibidi wound-healing chambers (Figure 2). The mature differentiated and polarized epithelial monolayers were infected with Hazara virus at two multiplicities of infection (MOI), 0.02 and 2.0 for 1 h, and wounds were then created at 0, i.e., immediately, or at 24 and 48 h post infection (hpi), allowing the cells to migrate and heal the wound for up to 52 h (Figures 2, 3). For the 0-hpi wound, we observed that the migration rates of virus-infected cells were similar to the control, i.e., untreated cells (Figure 3A). For the wounds created at 24 hpi, the rates of healing for cells infected with Hazara virus at MOI 2.0 were significantly suppressed at 30–52 h. By contrast, the MOI 0.02 resulted in significantly promoted migration between 24 and 30 h after creation of wounds, as compared to the untreated control (Figure 3B). When the wounds were created at 48 hpi, the migration rates of cells infected with Hazara virus at both MOI 0.02 and 2.0 and 0.02 were significantly inhibited around 24–28 h (Figure 3C); the control non-infected wounds were closed at that time and for this reason the study was not continued longer. In this set of experiments, no signs of Caco-2 cell death or apoptosis were seen.

To conclude, Hazara virus can perturb the normal physiology of epithelial cell monolayers and modulate their migration and wound-healing capacity in a dose- and time-dependent manner.

Individual and collective cell migration and restitution of epithelium are driven by actin cytoskeleton reorganization, primarily by dynamic polymerization of monomeric G-actin to F-actin (Ivanov et al., 2010; Friedl and Mayor, 2017) and are regulated by an ensemble of interacting proteins, including IQGAP1 (Karlsson et al., 2012; Hedman et al., 2015) and water fluxes via AQP (Karlsson et al., 2013), which facilitate F-actin formation. Using confocal imaging, we disclosed dramatic decreases in F-actin as well as redistribution of IQGAP1 and AQP6 in Caco-2 monolayers infected with Hazara virus at MOI 1 for 24 h, in clear contrast to control cells (Figures 4A–C). The distribution of these proteins in juxta-membrane regions went from a distinct uniform to a more disorganized and diffuse pattern; AQP6 re-localized to the cytoplasm; IQGAP1 formed cytoplasmic aggregates; and this was observed in more or less all cells in different fields of views. The quantification of fluorescence intensities of F-actin, IQGAP1 and AQP6 in juxta-membrane regions revealed a pronounced and significant decrease of intensity in virus-infected cells (Figures 4D–F), suggesting that cell-to-cell contacts and junctional associations are indeed affected. To further quantify, intensity profile plots across the cell monolayers were measured (Figures S1–S4). Here, non-infected controls displayed pronounced distinct assemblies of F-actin, IQGAP1 and AQP6 at cell-to-cell contacts, whereas the virus-infected cells had clearly more smooth profiles. Viral N was observed in the perinuclear region of infected cells (Figures 4A–C), which is in line with an earlier report (Andersson et al., 2004). Next, we investigated the effect of Hazara virus on the expression levels of IQGAP1 and AQP6 in epithelial cells with immunoblotting (Figure 5A) and further quantification of the density of the bands (Figures 5B,C). Here, the infection with Hazara virus at MOI 1 for 24 h caused a significant decrease in the expression level of either protein.

Thus, the Hazara virus infection resulted in less cytoskeletal F-actin, altered distribution and reduced expression of IQGAP1 and AQP6 appealing for perturbed organization of the cytoskeleton.

Since Hazara virus challenges resulted in modified epithelial migration (Figure 3) and reduced expression of several proteins important for cytoskeletal dynamics and homeostasis (Figures 4, 5), other parallel manipulations of cellular processes and programs might be of significance. We thus further aimed to identify partners of Hazara virus nucleocapsid protein (N) in the cellular proteome of human epithelia upon infection. To do so, epithelial monolayers were kept non-infected, or infected with Hazara virus at MOI 1 for 1 h, and at 24 hpi, total cell lysate samples were obtained and further immunoprecipitated using the antibodies against CCHFV/Hazara N. The N-enriched immune complexes, as well as total cell lysate samples were separated on SDS-PAGE and stained with Coomassie Blue (Figure 6). For the latter, no significant difference in protein patterns of infected and non-infected cells was observed (Figure 6, right panel). For the samples with N-enriched immune complexes, one additional protein band was reproducibly seen in the sample corresponding cells infected with Hazara virus (Figure 6, left panel).

Each gel line was cut into 10 slices, digested and analyzed by liquid chromatography and mass spectrometry (LC-MS/MS), where spectrum counts were used for quantitative comparisons of samples. To provide confidence, the experiments were repeated 4 times. The additional visible band detected in the sample corresponding cells infected with Hazara virus was identified as the viral N (Figure 6, left panel), which served as a control to confirm both infection and immunoprecipitation. Furthermore, 36 out of around 500 cellular proteins were typically identified in the N-enriched immune complexes and exhibited significant differences in affinity (Table 1).

Ten N-associated proteins were detected in virus-infected cells only, and 13 proteins displayed at least 2-fold increased binding to N-enriched immune complexes upon viral infection in comparison to the non-infected control (Table 1). These 23 proteins were considered as potential N partners in the Hazara infection. Twelve proteins displayed decreased binding to N-enriched immune complexes upon viral infection in comparison to non-infected control (Table 1). All 36 cellular proteins, typically identified in N-associated immune complexes and exhibited significant differences in affinity (Table 1), were included in the further bioinformatics analyses.

Bioinformatics analyses, using the search tool for the retrieval of interacting genes and proteins (STRING) and NCBI GO annotations by SCAFFOLD, were performed to investigate whether the identified N-associated cellular proteins could be placed into distinct functional classes. This revealed that they could be put in a groups being involved in: RNA and DNA processes, formation of membrane-bound vesicles, cell morphology, migration, differentiation, proliferation and apoptosis, metabolism, cellular defense and response to stress and pathogens, virus growth and chaperone activity (Table 1).

Independent analyses of network interactions, using STRING, disclosed that out of the 36 host proteins, which were typically identified in N-enriched immune complexes, at least 24 were connected through distinct types of actions (Figure 7), as previously being established experimentally and from curated databases and predicted by gene neighborhood and co-occurrence or by protein co-expressions and homologies (Figure S5). The distinct, large node cluster of N-associated proteins having a rich network of protein-protein interactions (Figure 7) contained proteins with chaperone activity (Table 1, depicted in blue): calnexin, mitochondrial Hsp60 and Hsp70, stress-induced phosphoprotein 1, elongation factor 1-δ, α-enolase, hsc70-interacting protein and two members of the protein disulfide isomerase family. Other proteins with a rich network were ribosomal receptor of activated protein C kinase 1 and cytoplasmic heterogeneous nuclear ribonucleoprotein Q (Figure 7). Further bioinformatics analyses revealed clusters of proteins involved in RNA and DNA processes (Table 1, noted in red, Figure S6A), in membrane-bound vesicle communication (Table 1, shown in violet, Figure S6B), and in key cellular processes, such as cell morphology, migration, differentiation, proliferation and apoptosis (Table 1, shown in green).

Thus, epithelial cellular protein networks are strongly perturbed by a Hazara virus infection, where the viral structural N associates and interacts with an array of cellular proteins.