key: cord-0262004-kb1c0cg7
authors: Barrass, Sarah V.; Pulkkinen, Lauri I. A.; Vapalahti, Olli; Kuivanen, Suvi H.; Anastasina, Maria; Happonen, Lotta; Butcher, Sarah J.
title: Proteome-wide cross-linking mass spectrometry to identify specific virus capsid-host interactions between tick-borne encephalitis virus and neuroblastoma cells
date: 2021-11-10
journal: bioRxiv
DOI: 10.1101/2021.10.29.464531
sha: 4a58c9fc79cca6ba0941ee2bcae9101d507138be
doc_id: 262004
cord_uid: kb1c0cg7

Virus-host protein-protein interactions are central to viral infection, but are challenging to identify and characterise, especially in complex systems involving intact viruses and cells. In this work, we demonstrate a proteome-wide approach to identify virus-host interactions using chemical cross-linking coupled with mass spectrometry. We adsorbed tick-borne encephalitis virus onto metabolically-stalled neuroblastoma cells, covalently cross-linked interacting virus-host proteins, and performed limited proteolysis to release primarily the surface-exposed proteins for identification by mass spectrometry. Using the intraviral protein cross-links as an internal control to assess cross-link confidence levels, we identified 22 high confidence unique intraviral cross-links and 59 high confidence unique virus-host protein-protein interactions. The identified host proteins were shown to interact with eight distinct sites on the outer surface of the virus. Notably, we identified an interaction between the substrate-binding domain of heat shock protein family A member 5, an entry receptor for four related flaviviruses, and the hinge region of the viral envelope protein. We also identified host proteins involved in endocytosis, cytoskeletal rearrangement, or located in the cytoskeleton, suggesting that entry mechanisms for tick-borne encephalitis virus could include both clathrin-mediated endocytosis and macropinocytosis. Additionally, cross-linking of the viral proteins showed that the capsid protein forms dimers within tick-borne encephalitis virus, as previously observed with purified C proteins for other flaviviruses. This method enables the identification and mapping of transient virus-host interactions, under near-physiological conditions, without the need for genetic manipulation. Author summary Tick-borne encephalitis virus is an important human pathogen that can cause severe infection often resulting in life-long neurological complications or even death. As with other viruses, it fully relies on the host cells, and any successful infection starts with interactions between the viral structural proteins and cellular surface proteins. Mapping these interactions is essential both for the fundamental understanding of viral entry mechanisms, and for guiding the design of new antiviral drugs and vaccines. Here, we stabilise the interactions between tick-borne encephalitis virus and human proteins by chemical cross-linking. We then detect the interactions using mass spectrometry and analyse the data to identify protein-protein complexes. We demonstrate that we can visualise the protein interaction interfaces by mapping the cross-linked sites onto the host and viral protein structures. We reveal that there are eight distinct sites on the outer surface of the viral envelope protein that interact with host. Using this approach, we mapped interactions between the tick-borne encephalitis virus envelope protein, and 59 host proteins, identifying a possible new virus receptor. These results highlight the potential of chemical cross-linking coupled with mass spectrometry to identify and map interactions between viral and host proteins.

mass spectrometry. We adsorbed tick-borne encephalitis virus onto metabolically-stalled neuroblastoma 23 cells, covalently cross-linked interacting virus-host proteins, and performed limited proteolysis to release 24 primarily the surface-exposed proteins for identification by mass spectrometry. Using the intraviral protein 25 cross-links as an internal control to assess cross-link confidence levels, we identified 22 high confidence 26 unique intraviral cross-links and 59 high confidence unique virus-host protein-protein interactions. The 27 identified host proteins were shown to interact with eight distinct sites on the outer surface of the virus. 28

Notably, we identified an interaction between the substrate-binding domain of heat shock protein family A 29 member 5, an entry receptor for four related flaviviruses, and the hinge region of the viral envelope 30 protein. We also identified host proteins involved in endocytosis, cytoskeletal rearrangement, or located in 31 the cytoskeleton, suggesting that entry mechanisms for tick-borne encephalitis virus could include both 32 clathrin-mediated endocytosis and macropinocytosis. Additionally, cross-linking of the viral proteins 33 showed that the capsid protein forms dimers within tick-borne encephalitis virus, as previously observed The three E proteins within each asymmetric unit are shown in blue, red, and yellow, and the prM protein is 96 shown in cyan. Symmetry axes are indicated by a black pentagon (five-fold), triangle (three-fold), and ellipse 97 (two-fold). 98

In this large-scale proteomics study, we used XL-MS to identify the interaction interfaces of PPIs between 99 TBEV and the surface of human neuroblastoma (SK-N-SH) cells. Here, the homobifunctional chemical cross-100 linker disuccinimidyl suberate (DSS) was used to covalently fix PPIs by cross-linking primary amine 101 containing residues (the side chain of lysine residues or the N-terminus of the protein). The finite length of 102 DSS (11.4 Å) imposed a maximum distance between cross-linked residues and was used to validate 103 intraviral crosslinks by measuring their distances on TBEV proteins with known structures or reliable 104 homology models. The final dataset was filtered using the intraviral cross-links as an internal control, 105 leading to the identification 59 unique high confidence interactions between the TBEV E protein and 106 cellular proteins. 107

Identification of cross-linked peptides 109 To identify interactions between the mature TBEV virion and host proteins, we incubated the virus with 110 metabolically-stalled neuroblastoma cells on ice, allowing for TBEV to bind to the cells, but preventing 111 The TBEV-host PPIs were then stabilized and fixed by chemical cross-linking  112 with DSS (Fig 2) . To reduce the sample complexity and search space during data analysis, we used limited 113 proteolysis to release primarily cell-surface associated host proteins. The released proteins were digested 114 to peptides and analysed by liquid chromatography tandem mass spectrometry (LC-MS/MS) followed by 115 label-free data dependent acquisition (DDA) quantitation to determine their relative abundance (S1 Table) . 116

Identified proteins were used to generate smaller, defined sets of sequences to use in the cross-linking data 117 analysis workflow. The released proteins were digested to peptides. 5. Peptides were analysed by LC-MS/MS and host and viral 123 proteins identified and quantified using label-free DDA. Identified host proteins were analysed to identify 124 cross-links between the host proteins and TBEV. 125

For cross-linking, we used four different cross-linker concentrations, in addition to a negative control 126 sample to which no cross-linker was added. Each condition was repeated in triplicate, and the experiment 127 repeated three times independently with different TBEV preparations and cell line passages, yielding 9 128 replicates per cross-linker concentration. The samples were initially analysed by immunoblotting of the 129 TBEV E and C Proteins (Fig 3) . The presence of higher molecular weight bands greater than 100 kDa in 130 samples treated with DSS confirms cross-linking (Fig 3) . The C protein has been shown to form antiparallel 131 dimers in the crystal and NMR structures of other flaviviruses [29-31]. We identified a band with a 132

[7] molecular weight corresponding to that of C protein dimers, indicating that the C protein dimerizes in TBEV 133 (Fig 3) . with 0.5 mM DSS; Lane 6-cross-linking with 1mM DSS. Higher molecular weight bands greater than 100 kDa 140 corresponding to the cross-linking of E and C to other proteins can be seen in lanes 2-5. Lower molecular 141 weight bands less than 50 kDa corresponding to the partial cleavage of the viral proteins during the limited 142 proteolysis step are also observed in all lanes. 143

Identification of cross-linked peptides is computationally challenging as all primary amine-primary amine 144 combinations in a given sequence database need to be considered. To reduce the search space for cross-145 linked peptide identification, proteins identified by DDA were probed for cross-links in batches (see 146 materials and methods). A total of 7167 spectral observations of cross-linked peptides were identified using 147 pLink2 at a false discovery rate (FDR) of 5%, excluding interfaces supported by cross-linked peptides 148 identified in the negative controls, which correspond to false positives likely arising from erroneous peptide 149 matches in the complex proteome background (S2 Table) [28]. Spectral observations of cross-links between 150

[8] two peptides within the same protein (intraprotein PPIs) accounted for the majority of the observations 151 (5589), compared to 1578 for those identified between two peptides from different proteins (interprotein 152 PPIs). Intraprotein cross-links have been consistently shown to make up a higher proportion of spectral 153 observations within cross-linked datasets, as two residues within the same protein are highly likely to be in 154 close physical proximity within the cell, leading to an increased cross-linking frequency [32, 33] . In total, 155 1697 different cross-linked interfaces were observed, and on average each interface was supported by 4.2 156 spectral observations. The cross-linked interfaces map to a network of 698 PPIs, consisting of 588 host 157 proteins and the 3 viral structural proteins. Overall, 66.3 % of the unique PPIs were attributed to 158 interactions between host proteins, 33.5 % to virus-host interactions and only 0.2 % to intraviral 159 interactions. We also identified intraprotein cross-links in 297 host proteins and the TBEV C and E proteins. 160

The confidence of the cross-linking dataset can be investigated by examining the spatial distances between 161 cross-linked residues for protein complexes where high-resolution structures or reliable homology models 162 are available. As intraviral protein interactions account for 62 % of the detected spectral observations, we 163 used the intraviral cross-links as an internal control to assess the confidence level for the dataset. 164

The published mature TBEV structure and homology models of the immature virus and C protein dimer 166 were used to accurately measure cross-link distances. The C-score of the homology models were -0.65 (C 167 protein), 2.00 (E protein) and 0.57 (prM protein) [34] [35] [36] . We measured the distance between cross-linked 168 residues and applied a maximum distance constraint of 30 Å between the lysine or N-terminus Cα (Table 1  169 and Figs 1 and 4) [37]. In total, 24 cross-links were mapped onto the viral structural proteins, and 22 fell 170 within the accepted distance range. Overall, 14 of the cross-links satisfied the distance constraint in both 171 the mature and immature TBEV structures, six were only acceptable in the mature structure and two in the 172 immature structure. Cross-links with acceptable distances in only the mature structure were identified by 173 2089 spectral observations compared to 31 for those only accepted in the immature structure, 174 demonstrating that the majority of the virus particles in the analysed samples were mature. 175 charge state of the cross-linked peptide is shown in blue. A) Representative spectrum of E protein cross-link 217 300-161 that are less than 30 Å and show a high number of spectral observations. B) Representative 218 spectrum for the cross-link 309-407, that has a cross-linking distance of 41.09 Å in the mature structure and 219 43.78 Å in the immature structure. C) Representative spectrum for the cross-link 309-408 that has a cross-220 linking distance of 39.54 Å in the mature structure and 41.45 in the immature structure. Cross-links with 221 distances > 30 Å show poor fragmentation and signal to noise ratio in the cross-linked spectral (B and C), 222 compared to cross-links with distances < 30 Å (A). 223

Based on the imposed distance constraints, we calculated that 91.7 % of the intraviral cross-linked 224 interfaces are identified with high confidence, at a distance threshold of <30 Å indicating that our approach 225 allows detection of specific intra-and inter-protein cross-links with high confidence. 226

Filtering the cross-linking dataset between the E-value, SVM score, number of spectral observations and the acceptable distance constraint 240 (<30 Å) as measured for our intraviral cross-links, to determine confidence parameters for the dataset. 241

[14]

Our data show a correlation between the measured distance of each unique intraviral cross-link and the 242 calculated SVM scores, but not the E-values (Fig. 6, S1 Fig) . Consequently, the E-value was not considered 243 as a suitable measure of confidence for this dataset. We observed that intraviral cross-links with distances 244 > 30 Å in both the mature and immature structures (indicated as red bars in Fig. 6 ) had the 3 rd and 4 th 245 highest mean SVM scores of the intraviral cross-links (Table 1 ). Furthermore, cross-links with higher 246 mean SVM scores were only identified by one spectral observation (Table 1) The number of cross-linked spectral observations for 101 randomly selected proteins in the dataset was 258 compared to the average spectral count of these proteins across all the samples in the DDA analyses (S1 259 File). The number of spectral observations for cross-linked peptide pairs for these same 101 selected 260 proteins was also compared to the number of lysine residues present within each protein (S1 File). No 261 correlation was observed in either of these cases; so, neither protein abundancy nor lysine content affect 262 the data content (S1 File). Therefore, the number of spectral observations was used in a non-biased 263 manner to assess the confidence of each unique cross-link to reduce the dataset for analysis further. 264

Here, we filtered the data in a stepwise manner, first on the PSM level using the SVM score and secondly 265 on the protein-protein interaction level using the number of spectral observations. A cross-linked interface 266 was considered to be of high confidence if it had an SVM score < 0.355 and was identified by ≥ 2 spectral 267 observations. Filtering the data in the reverse order would lead to inclusion of interactions only supported 268 by one high confidence spectral observation. The filtered cross-linking dataset consisted of 218 cross-linked 269 interfaces, 36.7 % of which were attributed to virus-host PPIs (S3 Table) . 270 interacting with assemble capsids. Only six of these residues shown in blue are also accessible on the outer 279 surface of the immature particle, as residue 336 is obscured by other E proteins and residue 251 by prM 280 (Fig 7) . The cross-linked lysine residues are distributed evenly across the surface of the E protein, showing 281 no preference for domains I, II or III (Fig 7) . 282 We performed String-database analysis to identify if any of the 59 E protein-interacting host proteins also 299 interact with each other (Fig 6) [40]. In total, 46 proteins were shown to interact with at least one other 300 protein, and nine interaction clusters were identified. This suggests that TBEV may interact with both 301 individual proteins and larger protein complexes. Gene ontology (GO) analysis (S4 Table) 

In this study, we present a chemical cross-linking proteomics approach to simultaneously identify TBEV-321 neuroblastoma cell PPIs and their interaction interfaces. We used metabolically-stalled cells to adsorb and 322 cross-link virus only to the cell surface, hoping to primarily enrich proteinaceous cell surface interactions. 323

The cross-linked proteins were released by limited proteolysis. Analysis of this highly complex protein 324 mixture by LC-MS/MS generated a large database of spectra containing four different peptide species with 325 the minority being cross-linked. In addition, cross-linked peptides are the least well-fragmented in the 326 database. In the next step, the pLink2 software compares the database of spectra with all of the possible 327 theoretical cross-linker reaction outcomes. This step is a clear bottle neck in the process as in our hands, 328 only 28 proteins could be analysed at a time, requiring multiple batch runs. Here, we optimised the analysis 329 workflow in order to extract the most significant virus-host PPIs from our complex data. Firstly, we reduced 330 the cross-linking search space by only analysing proteins identified by linear peptides in the samples. 331

Secondly, we have an internal validation control in the sample. We identified high-confidence intraviral 332 crosslinks using both the known and predicted three-dimensional structures of TBEV and the known length 333 of the chemical cross-linker. Then, we correlated the high confidence cross-links with the SVM score and 334 the quality of the spectra, allowing us to use an SVM score cut-off < 0.355 for the entire dataset. Finally, we 335 imposed a spectral count cut-off ≥ 2 to select for the most specific protein interactions. Using this method, 136) and the HSPA5 substrate binding domain residues 521 and 516 (Fig 9) . Therefore, the interaction 389 between HSPA5 and the E protein hinge region could constitute a unique binding interface, or be part of a 390 larger interface that also binds domain III consistent with other flavivirus studies [53-55,57]. We 391 hypothesize that HSPA5 could interact with both the hinge region of one E monomer and domain III of an 392 adjacent E monomer at the 3-fold axis, where the regions are in close spatial proximity (Fig 9) . Schematic representation of speculated HSPA5 binding region on the mature virus. The three E proteins 406 within each asymmetric unit are shown in blue, red, and yellow. The proposed HSPA5 binding site is shown 407 in green. 408

After the virus has attached to the cell surface, the next step is to enter the cells either through clathrin-409 mediated endocytosis or macropinocytosis, and we found evidence for both pathways being used. We Finally, cell debris was removed via centrifugation (16,000 x g, 5 min), the supernatant recovered, and the 459 samples prepared for mass spectrometry (Fig 2) . 

In total 874 proteins that were identified by 2 or more unique peptides, and an average spectral count of 510 ≥2 across all samples in the DDA data were probed for cross-links. Cross-links between the TBEV structural 511 proteins and host proteins were identified using the pLink2 software package [28] . In order to reduce the 512 search space for cross-link identification the data was analysed in 35 batches. The raw data dependent 513 acquisition data, and compiled FASTA file databases containing a total of 28 protein sequences, from 25 514 different host proteins and the 3 viral structural protein were used as the software input. Host protein 515 sequences were obtained from UniProt. The TBEV polyprotein sequence was obtained from GenBank 516 (accession: AWC08512.1), the C protein corresponds to residues 1-96, the M protein 206-280 and the E 517 protein 281-776. The following search parameters were used in the pLink2 software: Conventional cross-518 linking (Higher-energy C-trap dissociation (HCD)), precursor mass tolerance of 20 ppm; fragment mass 519 tolerance of 20 ppm; peptide length of 6-60; peptide mass of 350-6000 Da, up to 3 missed cleavage sites; 520 carbamidomethylation (C) was set to static modification; and oxidation (M) to variable modification. The 521 results were then filtered using a filtering tolerance of ±10 ppm and a separate FDR >5% at the peptide 522 spectrum matches level. 7716 cross-linked spectral observations were observed across all samples, 302 of 523 these were observed in the negative control samples. In total, 247 cross-linked spectral observations in the 524 0.1-1mM DSS samples corresponded to cross-linked interfaces also identified in negative control samples 525 and were excluded from further analysis. 526 [28] Homology modelling, structure visualization and measuring cross-link 527 distances 528 Homology models for the TBEV C protein, immature conformation of the E protein and the prM protein 529 were generated using the I-TASSER (Iterative Threading ASSEmbly Refinement) server [34] [35] [36] . A C-score 530 (confidence score for estimating the quality of predicted models by I-TASSER) is generated for each model 531 and can range from -5 to 2, with a higher value signifies a higher confidence and where a C-score > -1.5 is 532 considered good [34] [35] [36] . The C-protein homology model was generated using PDB accession 5OW2, and 533 the immature conformation of the E protein, and the prM protein using PDB accession 7L30 as the 534 template. The TBEV polyprotein sequence was obtained from GenBank (accession: AWC08512.1). The 535 sequence for the C protein was obtained from residues 1-96, the E protein residue 281-776, and the prM 536 residues 113-280 in the polyprotein. Generation of the assembled immature virus homology model, and 537 the C-protein dimer was performed in UCSF Chimera [79] . The homology model for the immature virus was 538 generated by superimposing the models for the immature E protein and prM protein onto the assembled 539 immature Spondweni virus (PDB accession 6ZQW) structure, using the MatchMaker function. The 540 homology model for the C protein dimer was generated by superimposing two models of the C protein 541 onto the Zika C protein dimer (PDB accession 5YGH), using MatchMaker. The atomic models were used to 542 position both intraprotein and interprotein cross-links by choosing the distance between Cα atoms in 543 Chimera [24, 25] . 544

Networking and clustering analysis 

Host-Directed Antiviral Therapy

Antivirals blocking entry of enteroviruses and therapeutic potential

Virus structure and structure-based antivirals. Curr Opin 572 Virol

Subunit Vaccine Design and Evaluation against Respiratory Syncytial Virus Using Reverse Vaccinology 575 Approach. Vaccines

Subunit Vaccines Against Emerging Pathogenic Human 577 Coronaviruses. Front Microbiol

A Single Point Mutation in the Rhinovirus 2B Protein Reduces the 579 Requirement for Phosphatidylinositol 4-Kinase Class III Beta in Viral Replication

Advances in high-throughput methods for the identification of virus 582 receptors

Two Critical Host Cell Receptors Which Facilitate Japanese Encephalitis Virus Entry Into Neurons

A Physical and Regulatory 587 Map of Host-Influenza Interactions Reveals Pathways in H1N1 Infection

Global landscape of 589 HIV-human protein complexes

An Unbiased 591 Screen for Human Cytomegalovirus Identifies Neuropilin-2 as a Central Viral Receptor

A SARS-CoV-2 protein 594 interaction map reveals targets for drug repurposing

Direct identification of ligand-596 receptor interactions on living cells and tissues

HATRIC-based 598 identification of receptors for orphan ligands

Chemical proteomics tracks virus 600 entry and uncovers NCAM1 as Zika virus receptor

Mapping the SARS-CoV-2-602

Host Protein-Protein Interactome by Affinity Purification Mass Spectrometry and Proximity-Dependent 603

Biotin Labeling: A Rational and Straightforward Route to Discover Host-Directed Anti-SARS-CoV-2 604

Global 606 Interactomics Uncovers Extensive Organellar Targeting by Zika Virus

Rapid determination of 609 quaternary protein structures in complex biological samples

Forms Serotype-and Local Environment-Dependent Interspecies Protein Complexes. mSystems

Isolation and molecular 614 characterization of a tick-borne encephalitis virus strain from a new tick-borne encephalitis focus with 615 severe cases in Bavaria, Germany. Ticks Tick-Borne Dis

Molecular epidemiology of tick-borne encephalitis virus inIxodes ricinus ticks in Lithuania

Tick-borne encephalitis: A review of epidemiology, clinical characteristics, and 620 management

EAN consensus review on 622 prevention, diagnosis and management of tick-borne encephalitis

Structure of tick-borne encephalitis 624 virus and its neutralization by a monoclonal antibody

The envelope glycoprotein from tick-borne 626 encephalitis virus at 2 Å resolution

Attenuation of Tick-Borne Encephalitis 628 Virus by Structure-Based Site-Specific Mutagenesis of a Putative Flavivirus Receptor Binding Site

Flavivirus maturation leads to 631 the formation of an occupied lipid pocket in the surface glycoproteins

A high-speed search engine pLink 2 with 633 systematic evaluation for proteome-scale identification of cross-linked peptides

Solution structure of dengue virus capsid protein 636 reveals another fold

West Nile Virus Core Protein. 638 Structure

Crystal Structure of the Capsid Protein from Zika Virus

Large Scale Chemical Cross-linking Mass Spectrometry Perspectives

Identification of cross-645 linked peptides from large sequence databases

I-TASSER: a unified platform for automated protein structure and 647 function prediction

The I-TASSER Suite: protein structure and function 649 prediction

I-TASSER server: new development for protein structure and function predictions. 651 Nucleic Acids Res

Distance restraints from 653 crosslinking mass spectrometry: Mining a molecular dynamics simulation database to evaluate lysine-lysine 654 distances: Evaluating Lysine-Lysine Distances by MD for XL-MS

Identification of cross-linked peptides from 656 complex samples

Chemical crosslinking and mass 658 spectrometry to elucidate the topology of integral membrane proteins

STRING v11: protein-661 protein association networks with increased coverage, supporting functional discovery in genome-wide 662 experimental datasets

High mobility group box 1 (HMGB1): a pivotal regulator of 664 hematopoietic malignancies

Cryo-EM 666 Reconstruction of Dengue Virus in Complex with the Carbohydrate Recognition Domain of DC-SIGN

Improving mass spectrometry analysis 669 of protein structures with arginine-selective chemical cross-linkers

Chemical cross-671 linking/mass spectrometry targeting acidic residues in proteins and protein complexes

C-terminal 674 fragment of human laminin-binding protein contains a receptor domain for Venezuelan equine encephalitis 675 and tick-borne encephalitis viruses

Human laminin 677 binding protein as a cell receptor for the tick-borne encephalitis virus. Zentralblatt Für Bakteriol

Integrin Mediates Internalization of Mammalian Reovirus

The Glycoprotein B Disintegrin-Like Domain Binds Beta 1 682 Integrin To Mediate Cytomegalovirus Entry

The Epstein-Barr virus BMRF-2 protein 684 facilitates virus attachment to oral epithelial cells

α5β1 integrin as a cellular coreceptor for human 686 parvovirus B19: requirement of functional activation of β1 integrin for viral entry

The ATPase ATP6V1A facilitates rabies virus 689 replication by promoting virion uncoating and interacting with the viral matrix protein

Ezrin is essential for the entry of Japanese encephalitis 692 virus into the human brain microvascular endothelial cells. Emerg Microbes Infect

Identification of heat shock protein A9 as a 695 Tembusu virus binding protein on DF-1 cells

Identification of GRP 78 (BiP) as a liver cell expressed 697 receptor element for dengue virus serotype 2. Arch Virol

A 699 functional interaction between GRP78 and Zika virus E protein. Sci Rep

Host Factor for Japanese Encephalitis Virus Entry and Replication in Mammalian Cells

Identification of Glucose-Regulated Protein 78 704 (GRP78) as a Receptor in BHK-21 Cells for Duck Tembusu Virus Infection

Heat shock proteins bind calcitonin

ERdj5 is the ER reductase that catalyzes 708 the removal of non-native disulfides and correct folding of the LDL receptor

Unraveling the role of 711

KIAA1199, a novel endoplasmic reticulum protein, in cancer cell migration

LOXL2 drives epithelial-714 mesenchymal transition via activation of IRE1-XBP1 signalling pathway. Sci Rep

GRP78 clustering at the cell 716 surface of neurons transduces the action of exogenous alpha-synuclein

Regulation of tissue 719 factor--mediated initiation of the coagulation cascade by cell surface grp78

Endothelial Cells: Role for Grp78/BiP in T-Cadherin-Dependent Cell Survival

Molecular Chaperone BiP Interacts with Borna 726 Disease Virus Glycoprotein at the Cell Surface

Recent advances in the structural and mechanistic aspects of Hsp70 728 molecular chaperones

Zika virus envelope -heat shock protein A5 (GRP78) binding site prediction. J 730 Biomol Struct Dyn

Dissecting the Cell Entry Pathway of Dengue Virus by Single-Particle Tracking in Living Cells. Farzan M, 733 editor

Characterization of dengue virus entry into HepG2 cells

Characterization of Langat virus 737 antigenic determinants defined by monoclonal antibodies to E, NS1 and preM and identification of a 738 protective, non-neutralizing preM-specific monoclonal antibody

Numerical 740 Compression Schemes for Proteomics Mass Spectrometry Data

openBIS: a flexible 745 framework for managing and analyzing complex data in biology research

A method for reducing the time required to match protein sequences with 748 tandem mass spectra

Open Mass Spectrometry 750 Search Algorithm

Comet: An open-source MS/MS sequence database search tool. 752 PROTEOMICS

Empirical Statistical Model To Estimate the Accuracy 754 of Peptide Identifications Made by MS/MS and Database Search

Protein 756 Identification False Discovery Rates for Very Large Proteomics Data Sets Generated by Tandem Mass 757 Spectrometry

UCSF Chimera?A 759 visualization system for exploratory research and analysis

A Software Environment for Integrated Models of Biomolecular Interaction 761 Networks

GOnet: a tool for interactive Gene Ontology analysis

Supporting information captions 765 S1 Table: DDA analysis of viral and host protein in all samples (XLSX)

Cross-linking dataset Identified cross-linked peptides, corresponding proteins, E-values and SVM 767 scores for each cross-linked spectral observation (XLSX)

Table) filtered to remove spectral 769 observations corresponding to unique cross-links identified by less than 2 spectral observations with

GO analysis of host proteins GO analysis of the 59 host proteins shown to interact with the 772 surface of TBEV in the filtered cross-linking dataset

Correlation between number of cross-linked spectral observations, protein abundance and 774 protein lysine content Table showing the number of cross-linked spectral observations, the average protein 775 abundance calculated in the DDA analysis (S1 Table) and the number of lysines for 101 randomly selected 776 proteins. Scatter graphs show no correlation between the number of cross-linked spectral observations and 777 the number of lysines for the proteins

Box and whisker plot of the E-value for the intraviral cross-links, plotted against the distance 780 between the cross-linked residues for each cross-link. The smallest distance out of the mature and 781 immature calculated distances, and the C protein dimer or monomer distances is plotted

All peptide analyses were performed on a Q Exactive HFX mass spectrometer (Thermo Scientific) connected 480 to an EASY-nLC 1200 ultra-high-performance liquid chromatography system (Thermo Scientific). The 481 peptides were loaded onto an Acclaim PepMap 100 (ID 75μm x 2 cm, 3 μm, 100 Å) pre-column and 482 separated on an EASY-Spray column (Thermo Scientific; ID 75 μm × 25 cm, column temperature 45 °C) 483 operated at a constant pressure of 800 bar. A linear gradient from 4% to 45% of 0.1% formic acid in 80% 484 acetonitrile was run for 50 min at a flow rate of 300 nl/min One full MS scan (resolution 60,000@200 m/z; 485 mass range 350 to 1600 m/z) was followed by MS/MS scans (resolution 15,000@200 m/z) of the 15 most 486 abundant ion signals. The precursor ions were isolated with 2 m/z isolation width and fragmented using 487 higher-energy collisional-induced dissociation at a normalized collision energy of 30. Charge state screening 488 was enabled, and precursors with an unknown charge state and singly charged ions were excluded. The 489 dynamic exclusion window was set to 15 s and limited to 300 entries. The automatic gain control was set to 490 3 x 10 6 for MS and 1 x 10 5 for MS/MS with ion accumulation times of 110 and 60 ms, respectively. The 491 intensity threshold for precursor ion selection was set to 1.7 x10 4 . 492 The authors declare that they have no conflict of interest. 566