Cloning
Entry clones containing genes of interest were obtained from the ORFeome collection (ORFeome and MGC Libraries; Genome Biology Unit supported by HiLIFE and the Faculty of Medicine, University of Helsinki, and Biocenter Finland). The gene of interest used for generating the stable cell lines was fused to MAC-tag-C (Addgene, Plasmid #108077) destination vector using Gateway cloning techniques. The plasmids used in this project are available from Addgene. The gene of interest used for protein interaction validation was fused to a modified pDEST40 vector (with a 3×V5 C-terminal tag) (Varjosalo et al, 2008) using Gateway cloning techniques.
Point mutations of SARS-CoV-2 variants were created by performing site directed mutagenesis to plasmids using Q5 high-Fidelity DNA polymerase (NEB, M0491L). Primers were obtained from Eurofins Genomics. Mutations of interest were designed based on GISAID database (GISAID.ORG) and WHO classification of Variant of concern/interest including Alpha (B.1.1.7 and Q lineages), Beta (B.1.351 and descendent lineages), Gamma (P.1 and descendent lineages), Delta (B.1.617.2 and AY lineages), and Omicron (B.1.1.529, BA.1, lineages). Final selection of the mutation of interest (total of 50 construct) was based on the significant occurrence of mutation within variant strains.
Cell culture
The Flp-In™ 293 T-REx cell line (Invitrogen, R78007), HEK293 cell line (ATCC® CRL-1573™), and U2-OS (ATCC®, HTB-96™) were maintained under the manufacturer-recommended conditions. Transfection and generation of stable cell lines were performed as previously described (Liu et al, 2020). Briefly, Flp-In™ 293 T-REx cells were co-transfected with the MAC-tagged bait or GFP construct and pOG44 vector (V600520, Invitrogen) using FuGENE 6 transfection reagent (Promega, Wisconsin, USA). Forty-eight hours post-transfection, cells were selected for 3 weeks in medium containing hygromycin B (100 μg/ml; Invitrogen). The positive clones, containing stable isogenic MAC-tagged baits, were further expanded to 80% confluence in 10 × 150-mm cell culture plates (CELLSTAR, Greiner) in complete culture medium. Cells from five plates were used for one biological replicate, and to induce the expression of protein of interest 1 μg/ml tetracycline (Sigma-Aldrich, T3383-25G) was added, together with 50 μM biotin (Thermo Fisher, 29129) for 24 h before harvesting.
Virus preparation
SARS-CoV-2 variants were isolated from nasopharyngeal samples in VeroE6 or VeroE6-TMPRSS2-H10 (Rusanen et al. 2021) cells: FIN34-21 (Alpha variant; GenBank ON532062.1), FIN32-21 (B.1.351, Beta variant; OK448476.1 and EPI_ISL_3471851), FIN37-21 (B.1.617.2, Delta variant, OK626882.1 and EPI_ISL_2557176) and Omicron variants FIN56-21 (BA.1.1 variant, EPI_ISL_8586102) and FIN61-22 (BA.5 variant, OP435368 and EPI_ISL_13118918). Viruses were further amplified in VeroE6-TMPRSS2-H10 cells in DMEM supplemented with 2% fetal calf serum, 2mM L-glutamine, and 2 mM penicillin-streptomycin. Virus stocks were tittered in VeroE6-TMPRSS2-H10 cells by median Tissue Culture Infectious Dose (TCID50) assay as described before (Belik et al. 2023; Ardito et al. 2017).
Infection experiments
For infection, VeroE6-TMPRSS2-H10 cells were seeded on the previous day (3x106 cells per 10 cm2 dish, 0.4 x106 cells per 6-well plate well, and 0.075 x106 cells per 24-well plate well) in DMEM supplemented with 10% fetal calf serum, 2mM L-glutamine, and 2 mM penicillin-streptomycin. At the time of infection, the media was changed to DMEM supplemented with 2% fetal calf serum, 2mM L-glutamine, and 2 mM penicillin-streptomycin, and cells were kept in this media for the rest of the time. Cells were infected with SARS-CoV-2 variants in half of the designated, final medium volume. Following infection, the cells were incubated in +37°C and 5% CO2 for 1 hour, after which the remaining half volume of the infection medium was added. Note, that the inoculated virus was not removed. Depending on the experiment and the virus, the final titers were 20 000 – 70 000 TCID50/ml. In proteome and RNA analyses, the final titers were: 70 000 TCID50/ml Fin-34, 58 000 TCID50/ml Fin-37, and 25 000 TCID50/ml for Fin-32, Fin-56 and Fin-61. For immunofluorescent analyses the final titers were 56 000, 46 400 and 20 000 TCID50/ml, respectively. Protein and RNA samples were collected and cells for immunofluorescence were fixed at time points of 1hpi, 8hpi, 16hpi and 36hpi. Samples from uninfected cells were collected as controls.
Sample preparation from the infected samples
For protein samples, the cells were infected with 3 – 8 x 105 TCID50 of variant per 10 cm2 plate. At the time of sample collection, the plates were kept on cold plate, the medium was removed, cells were washed once with ice-cold TBS, 0.5ml ice-cold freshly-made 8M urea solution (in 64mM NH₄HCO₃ supplemented with Phosphatase Inhibitor Cocktail (Roche)) was added per 10 cm2 dish, and the cells were scraped off the plate. Cell lysates were kept on ice and stored at -80°C until analysis.
For RNA samples, the cells were infected with 0.5 – 1.4 x 105 TCID50 of variant per 6-well plate well. At the time of sample collection, the plates were kept on cold plate, the medium was removed, cells were washed once with ice-cold PBS, 0.3 ml RA1 buffer (Macherey-Nagel) with 20mM DTT was added per 6-well plate well, and the cell lysate was suspended by careful pipetting. Cell lysates were kept on ice and stored at -80°C until analysis.
Infection rate estimation with immunofluorescence
For immunofluorescence, the cells were infected with 0.1 – 0.28 x 105TCID50 of variant per 24-well plate well. The cells were fixed with 4% formaldehyde for 1 hour, the plates were washed with PBS, and stored in PBS at +4°C until analysis. The fixed cells were blocked and permeabilizated with 0.5% bovine serum albumin (BSA, Sigma-Aldrich) and 0.1% Triton X-100 (Sigma-Aldrich) in phosphate-buffered saline (PBS, Gibco) for 30 minutes at room temperature. Subsequently, primary antibodies in-house SARS-CoV-2 rabbit anti-N (Maljanen et al., manuscript) and rabbit anti-S1 antibodies (Kolehmainen et al. 2023), diluted in 3% BSA in PBS, were added for 1 hour at room temperature. The cells were washed thrice for 10 minutes with PBS. Secondary antibodies (Goat α-rabbit IgG (H+L) AlexaFluor Plus 488, Thermo Fisher Scientific) and 4′,6-diamidino-2-phenylindole (DAPI, Life Technologies), both diluted in 3% BSA in PBS, were applied to the cells for one hour at room temperature in dark. The cells were washed thrice for 10 minutes in 3% BSA in PBS. Subsequent imaging was performed utilizing the EVOS FL Auto imaging system (Life Technologies). The images were processed in Fiji, a common, open-source distribution of ImageJ (Schindelin et al. 2012). To estimate the infection percentage, the nuclei were segmented using the StarDist2D plugin in Fiji (Schmidt et al. 2018). A binary classifier was trained using the Trainable Weka Segmentation plugin (Arganda-Carreras et al. 2017) in Fiji to identify which cells were infected and which were not. Roughly 20% of the images taken in the first repetition were used for training the algorithm and the remaining 80% of the images from same repetition were used for validation. The resulting algorithm was then used on subsequent repetitions and image sets.
Quantitative proteomics and phosphoproteomics sample preparation
Viral infected cell samples were lysed in lysis buffer (8M Urea, 50 mM NH4HCO3 supplemented with 1X phosphatase and protease inhibitors cocktail (Sigma-Aldrich, P2745 and P8340) on ice. The lysate was cleared by centrifugation at 16,000 × g for 10 min at 4 ˚C. The protein concentration was determined using a BCA protein assay kit (Thermo Fisher). Equal amounts of protein (250 µg) were obtained for all samples. Proteins were reduced with Tris(2-carboxyethyl) phosphine (TCEP; Sigma Aldrich), and alkylated with iodoacetamide. Samples were diluted 4-fold (to less than 2 M urea) with 50 mM ammonium bicarbonate (AMBIC; Sigma-Aldrich), and digested with Sequencing Grade Modified Trypsin (Promega) in approximate ratio of 1:50 to 1:20 (w/w) at 37°C for 16 hours. Finally, the peptide samples were desalted with C18 Macrospin columns (Nest Group). Desalted peptides were divided for whole proteome analysis and following phospho-proteome analysis.
For phosphopeptide enrichment, desalted peptide mixtures were firstly loaded into a stage-tip packed with Ti4-IMAC microspheres which were prewashed with loading buffer (80% acetonitrile/6% trifluoroacetic acid). The Ti4-IMAC microspheres with enriched phosphopeptides were collected by centrifugation. To remove nonspecifically absorbed peptides, the Ti4-IMAC microspheres were washed with 50% acetonitrile/6% trifluoroacetic acid/200mM NaCl and 50% acetonitrile/0.1% trifluoroacetic acid sequentially. To elute the enriched phosphopeptides, elution buffer containing 10% NH4OH was added and the enriched phosphopeptides were eluted with centrifugation. The supernatant containing phosphopeptides was collected and lyophilized for LC-MS/MS analysis.
Proximity labeling
For BioID-MS approach, cell pellets were thawed in 3 ml of ice-cold lysis buffer (0.5% IGEPAL, 50 mM HEPES (pH 8.0), 150 mM NaCl, 50 mM NaF, 1.5 mM NaVO3, 5 mM EDTA, 0.1% SDS, 0.5 mM PMSF, and protease inhibitors), and lysates were sonicated and treated with Benzonase® Nuclease (Santa Cruz Biotechnology, sc-202391).
Lysate was then centrifuged, and clear solution containing target proteins was obtained. Lysate was subjected to a one-step purification via Strep-Tactin® Sepharose® resin (IBA). A detailed description of the method used here can be found in a previous protocol (Liu et al, 2020).
MS Sample preparation
Purified protein lysis sample was prepared further to for LC-MS. To break disulfide bonds, the samples were incubated with 5 mM TCEP (Tris(2-carboxyethol) phosphine hydrochloride salt, #C4706 Sigma-Aldrich, USA) at 37 °C for 20 min. Samples were let cool to room temperature and alkylated with 10 mM iodoacetamide (#57670 Fluka, Sigma-Aldrich, USA) at room temperature for 20 min. Samples were trypsin-digested with 1.5 µg Sequencing Grade Modified Trypsin (V5111, Promega) at 37°C for 16 hours. After digestion peptides were quenched with 10% trifluoroacetic acid (TFA) and finally, the desalted with BioPureSPN MINI C18 columns (The nest Group, inc.) A detailed description of the method used here can be found in a previous protocol (Liu, Salokas et al., 2020). The dried peptides were reconstituted in 30 µl Buffer A (0.1% (vol/vol) TFA and 1% (vol/vol) acetonitrile (ACN) in HPLC water). Samples were further diluted 1+19 µl with HPLC water containing 0.1 vol/vol% formic acid. The manufacturer's instructions were followed to load into Evotips (Evosep).
Liquid chromatography-mass spectrometry (LC-MS) analysis
The desalted samples underwent analysis using the Evosep One liquid chromatography system, which was coupled to a hybrid trapped ion mobility quadrupole TOF mass spectrometer (Bruker timsTOF Pro) through a CaptiveSpray nano-electrospray ion source. Peptide separation was achieved using an 8 cm × 150 µm column packed with 1.5 µm C18 beads (EV1109, Evosep) following the 60-sample-per-day method, which employed a 21-minute gradient time. Mobile phases A and B consisted of 0.1% formic acid in water and 0.1% formic acid in acetonitrile, respectively. For BioID samples, the mass spectrometry (MS) analysis was conducted in positive-ion mode using a data-dependent acquisition (DDA) approach in PASEF (Parallel Accumulation Serial Fragmentation) mode, utilizing the DDA-PASEF-short_gradient_0.5s-cycletime method.
The DIA MS analysis was performed in the positive-ion mode with DIA-PASEF method (Meier et al. 2020) with sample optimized data independent analysis (DIA) scan parameters. We performed DDA in PASEF mode from a pooled sample to be able to adjust DIA-PASEF parameters optimally to each sample type. To perform sample specific DIA-PASEF parameter adjustment the default DIA-short-gradient acquisition methods was adjusted based on the sample specific DDA-PASEF run with the software “tims Control” (Bruker Daltonics). The following parameters were modified for each sample type: phoshory enrichment; Mass Range 433 to 1684 Da, Mobility Range 0.85 to 1.3 1/K0, Cycle time estimate 1.80 seconds and for total proteome; Mass Range 370-1370 Da, Mobility Range 0.85 to 1.30 1/K0, with cycle time estimate 1.27 seconds. The ion mobility windows were set to best match the ion cloud density from the sample type specific DDA-runs.
Following the analysis, the raw data obtained for both total proteome and Phosphoproteomics underwent processing using DIA-NN (version 1.8.1) software. For Total Proteome DIA-NN, the parameters included a maximum of 1 missed cleavage, up to 3 variable modifications, N-terminal methionine excision, carbamidomethylation of cysteine residues, oxidation of methionine, acetylation of the N-terminus, peptide length ranging from 7 to 30 amino acids, precursor charge ranging from 2 to 4, precursor m/z range of 300-1600, fragment ion m/z range of 100-1700, and mass accuracy of 15 ppm, with MS1 accuracy also set at 15 ppm. For Phosphoproteomics DIA-NN, the parameters were similar, with 1 missed cleavage allowed, a maximum of 3 variable modifications, N-terminal methionine excision, carbamidomethylation of cysteine residues, exclusion of oxidation on methionine residues, acetylation permitted at the N-terminus, peptide length ranging from 7 to 30 amino acids, precursor charge ranging from 2 to 4, precursor m/z range of 300-1600, fragment ion m/z range of 100-1700, and a mass accuracy of 15 ppm, with MS1 accuracy also set at 15 ppm.
BioID Raw data were processed with FragPipe v17.1 utilizing MSFragger (Yu et al. 2020) against reviewed human entries of the UniProtKB database (downloaded 8.3.2022) coupled with SARS-CoV-2 protein information derived from GISAID. Carbamidomethylation of cysteine residues was used as static modification. Amino terminal acetylation and oxidation of methionine were used as the dynamic modification. Biotinylation of lysine and N-termini were set as variable modifications. Trypsin was selected as enzyme, and maximum of two missed cleavages were allowed. Both instrument and label-free quantification parameters were left to default settings. Final results from these steps are Spectral Counts (SC) values from peptides with FDR < 0.01 from Philosopher.
Identification of high-confidence protein-protein interactions
The web tool (http://proteomics.fi/) incorporated with Significance Analysis of INTeractome (SAINT) -express version 3.6.3 (Teo et al. 2014) was used as a statistical approach for identification of specific high-confidence interactions from BioID-MS data. High-confidence interactions (HCIs) were defined by an estimated protein-level Bayesian FDR (BFDR) of ≤ 0.05. Furthermore, we used our own in-house contaminant GFP library for Proximity Depended Biotinylation the with a cutoff frequency of ≥ 50%, except for an average spectral count fold change ≥ 3 to remove non-specific interactors. Most common Mass-spectrometry contaminants were removed from results. Results are bait normalized.
MS-microscopy analyses
For any bait of interest, the averaged peptide-spectrum match (PSM) values of each prey protein were calculated and uploaded to the web tool (http://proteomics.fi/) to calculate their subcellular distribution.
Immunostaining and fluorescence microscopy
Transfected U2OS U2-OS (ATCC®, HTB-96™) cells were cultivated in 96well optical bottom plate (Brooks, MGB096-1-2-LG-L). After fixation in 4% paraformaldehyde in PBS and permeabilization in Triton-x100, cells were incubated with primary (anti-HA IgG1, BioLegends 901502) antibodies. Next, cells were washed twice with PBS and incubated with PhenoVue Fluor 488 - Donkey Anti-Mouse Antibody Cross-Adsorbed (PerkinElmer, 2DXM488C1); PhenoVue Fluor 568 - Phalloidin (PerkinElmer, CP25681); PhenoVue Fluor 647 - Concanavalin A (PerkinElmer, CP96471); DAPI (Santa Cruz, 28718-90-3). After staining cells were washed twice and imaged in PBS.
Imaging was performed using a Molecular Devices Nano high-content screening microscope, using Nikon 20x/0.45 S Plan Fluor ELWD, WD 8.2-6.9mm (pixel size 0.328 um) objective. Following filter sets were used for multicolor acquisition: DAPI (Ex: 377/50 nm, Em: 447/60 nm, Dichroic: 409 nm), GFP (Ex: 472/30 nm, Em: 520/35 nm, Dichroic: 495 nm), Texas Red (Ex: 562/40 nm, Em: 624/40 nm, Dichroic: 593 nm), Cy5 (Ex: 628/40 nm, Em: 692/40 nm, Dichroic: 660 nm). Images we stored as TIF and processed (linear intensity scaling, pseudocoloring, composite panels assembly) using custom-made python script.
Data Preprocessing:
The experimental data underwent bait normalization as a preliminary step. Following bait normalization, missing values in the data were imputed using the QRILC (Quantile Regression-based Imputation using Locally Constant functions) method using the `imputeLCMD` R package and R (version 4.3.1). Furthermore, median normalization was conducted on the data as a normalization step prior to analysis.
Correlation Analysis:
A differential abundance analysis was performed to compare protein abundance levels across sample groups. Log2 fold change and p-value were calculated using python (version 3.9.7) and ttest_ind from scipy.stats python package based on the imputed and normalized intensity data. The p-values were corrected using Benjamini-Hochberg correction method from the `multitest` module of the `statsmodels.stats` package. Figures were generated using plotly python library.
Gene Ontology (GO) enrichment analysis
We employed the Database for Annotation, Visualization, and Integrated Discovery (DAVID) to unravel the functional intricacies of the proteome dataset. The primary goal was to uncover enriched Gene Ontology (GO) terms and pathways that shed light on the molecular context of the identified proteins. Our investigation encompassed three key aspects of gene ontology: Biological Process (BP), Cellular Component (CC), and Molecular Function (MF). DAVID discerned GO terms that exhibited significant enrichment within our protein dataset. Simultaneously, we delved into the intricate web of cellular pathways by applying the KEGG and Reactome pathway analyses. (Huang, Sherman, and Lempicki 2009a; 2009b)
Structural analysis of mutation sites
In order to present the three-dimensional (3D) structure of our proteins of interest, PyMol version 2.5 (Schrödinger LLC, New York, NY, USA) was used. The crystal structure of the SARS-CoV-2 proteins was obtained from the Protein Data Bank (PDB), while de novo models were taken from the AlphaFold protein folding prediction database and Swiss-Prot database. Figures of the protein structure were generated using the PyMol software. database(Jumper et al. 2021; Varadi et al. 2022)