Simultaneous membrane and RNA binding by tick-borne encephalitis virus capsid protein

Tick-borne encephalitis virus is an enveloped, pathogenic, RNA virus in the family Flaviviridae, genus Flavivirus. Viral particles are formed when the nucleocapsid, consisting of an RNA genome and multiple copies of the capsid protein, buds through the endoplasmic reticulum membrane and acquires the viral envelope and the associated proteins. The coordination of the nucleocapsid components to the sites of assembly and budding are poorly understood. Here, we investigate the interactions of the wild-type and truncated capsid proteins with membranes with biophysical methods and model membrane systems. We show that capsid protein initially binds membranes via electrostatic interactions with negatively-charged lipids, which is followed by membrane insertion. Additionally, we show that membrane-bound capsid protein can recruit viral genomic RNA. We confirm the biological relevance of the biophysical findings by using mass spectrometry to show that purified virions contain negatively-charged lipids. Our results suggest that nucleocapsid assembly is coordinated by negatively-charged membrane patches on the endoplasmic reticulum and that the capsid protein mediates direct contacts between the nucleocapsid and the membrane.


Author summary
Tick-borne encephalitis virus (TBEV) causes a life-threatening neurological disease. The incidence of the disease is increasing despite the availability of efficient vaccines. The virus particles contain a nucleocapsid consisting of the viral RNA genome and multiple copies of the capsid protein. The nucleocapsid is surrounded by an envelope derived from host membranes and two viral surface protein species are embedded into it. The assembly of TBEV virions is poorly understood despite its medical significance. Here, we have used a large combination of biophysical methods to study the interactions of the capsid protein As the C protein has membrane affinity, and the assembly process occurs at the ER, it is likely that specific ER lipids are involved in the process. Budding involves generating extreme membrane curvature which is only possible in the presence of suitable lipids. Flavivirus infection leads to a significant increase in curvature-promoting lipids in the host, and increases the curvature of the ER membrane [43][44][45][46][47][48][49][50]. Previous work has mainly concentrated on fluorescence microscopy-based characterization of C protein localization in infected cells, experiments conducted with C-derived peptides, and indirect characterization of membrane insertion [40,41,[51][52][53]. Here, we have probed the role of lipids in the NC assembly by using recombinant TBEV C and various model lipid systems to biophysically characterize the interaction of the C protein with membranes. We show that the C protein initially binds to lipids with negatively-charged head groups via electrostatic interactions which is then followed by membrane insertion. We confirm the physiological relevance of the experimental system by showing that the C protein can recruit genomic RNA whilst membrane-bound in the presence or absence of the 17 N-terminal residues. Furthermore, we characterise the TBEV lipidome with ultra-high performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UHPLC-QTOF-MS) and show that the TBEV virion contain lipids with negativelycharged head groups. Our results indicate that the C protein may be recruited to the ER via positively charged lipid head groups, helping to nucleate the assembly of the NC and acquisition of the membrane through budding into the ER.

Purified C protein binds to lipids with negatively-charged head groups
The homology model of the full-length C protein from the Kuutsalo-14 strain was predicted using the I-TASSER server by using the structure of the truncated C protein (PDB: 7YWQ) as a template (S1B Fig) [34,[54][55][56][57]. The C-score of the homology model was -0.69 (on a scale of -5 to 2), indicating high confidence structure prediction [54][55][56]. The homology model had an overall similar fold to the truncated C protein, with the N-terminus forming an additional α helix (S1B Fig). Electrostatic surface potential calculation showed that in addition to the positively-charged patch on the α 4 -α 4 interface, the N-terminus is also positively charged (S1C Fig).
The homology model was used to guide C protein mutagenesis (S1B Fig) and both the full-length C and the truncated C (C 18-93 , containing residues 18-93) were expressed in E. coli. After immobilized metal affinity chromatography and ion-exchange chromatography, both proteins eluted as pure dimers in size-exclusion chromatography (SEC) confirmed by SDS-PAGE and immunoblotting (S2 Fig). The A 260 /A 230 ratios of the purified proteins were 0.61±0.02 and 0.55±0.3 for C and C 18-93 respectively indicating no significant nucleic acid contamination (average of three technical replicates) [58].
To determine if the full-length C protein bound to membranes and if lipid binding was dependent on the type of lipid head groups, we used liposome co-sedimentation. First, we used liposomes containing either brain total lipid extract (BTLE), or a mixture of the uncharged lipids 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE). Full-length C protein was found to bind and co-sediment with BTLE liposomes, but not with DOPC-DOPE (Fig 1A and 1B). Next, we supplemented the DOPC-DOPE liposomes with individual glycerophospholipids, sphingolipids or cholesterol (Chol) representing different specific lipid properties. Incorporation of sphingomyelin (SM), phosphatidylinositol (PI), or galactocerebrosides (GalCer) showed that lipids with uncharged or glycosylated head groups did not markedly increase full-length C protein binding, neither did Chol. However, clear binding was detected after incorporation of the negatively-charged 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (POPS) or the highly negatively-charged L-α-phosphatidylinositol-4,5-bisphosphate (PI(4,5)P 2 ) (Fig 1A and  1B). This indicated that full-length C bound to negatively-charged membranes without a preference for a specific head group. Furthermore, PI(4,5)P 2 that is 3 times more charged than POPS resulted in increased binding (Fig 1A and 1B). To further confirm the role of charge, DOPC-DOPE liposomes were supplemented with increasing concentrations of POPS, increasing the negative charge of the liposomes proportionally, and co-sedimentation was then assessed. The fraction of co-sedimenting full-length C protein increased with increasing POPS concentrations, with all protein being bound at 40% POPS (Fig 1C and 1D). Thus, 40% POPS was used for further experiments.
We confirmed the results of the co-sedimentation with liposome flotation assays. Fulllength C protein co-migrated with liposomes containing 40% POPS. The majority of the C shows densitometric quantitation of the co-sedimenting fraction from the pellet. PC/PE refers to pure DOPC/DOPE liposomes, BTLE to pure BTLE liposomes, a dash to the no liposome control and the other labels to DOPC/DOPE liposomes supplemented POPS, Chol, SM, PI, PI(4,5)P 2 or GalCer. C, D Full-length C protein co-sedimentation with liposomes containing varying amounts of POPS. (C) shows representative SDS-PAGE bands where S indicates the supernatant and P the pellet with the POPS percentage indicated above the lanes. (D) shows densitometric quantitation of the co-sedimenting fraction from the pellet with the w/w percentage of POPS in the liposomes is indicated on the Y axis. Data information: In (B) and (D), data are shown as the averages of three replicates with the error bars representing the standard deviation (s.d.

PLOS PATHOGENS
Membrane and RNA binding by tick-borne encephalitis virus capsid protein protein was detected in the fraction containing the fluorescently-labelled lipids based on SDS-PAGE analysis and fluorescence measurements (S3 Fig). Contrastingly, in experiments with liposomes containing no negatively-charged lipids, the C protein did not co-migrate and the majority of the protein was detected in the fractions without liposomes (S3 Fig). Similar results were obtained in control experiments without liposomes (S3 Fig). This confirmed that C protein binds liposomes with charged lipid headgroups, rather than precipitating in the cosedimentation assay.

Initial C protein recruitment to the membrane is of an electrostatic nature
To further characterise the binding of C protein to membranes, supported lipid bilayers (SLBs) on a quartz crystal microbalance with dissipation monitoring (QCM-D) measurements were used. SLBs were generated either from pure 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) or a mixture of POPC and POPS (60:40% w/w) (Fig 2A). QCM-D measures the change in frequency (ΔF) of an oscillating quartz crystal which is related to the mass adsorbed to its surface. It also monitors the energy dissipation (ΔD) representing how rigid the adsorbed material is, a high dissipation corresponding to a more viscoelastic (soft) film. During formation of SLBs the adsorption of liposomes and subsequent fusion into a two-dimensional rigid SLB can be scored using these parameters (Fig 2A) [59]. After the formation of the SLB and the equilibration of the system with Tris-buffered saline (TBS), the ΔF and ΔD values were zeroed, and the C protein was injected. The C protein injection caused the ΔF to decrease to −23.9 Hz on average (s.d. 2.2 Hz) in POPC/POPS experiments indicating a clear binding of additional mass on to the SLB (Fig 2B and 2C). However, repeating the experiment in the presence of pure POPC SLBs, resulted in a significantly smaller ΔF after C protein injection (on average 0.9 Hz, s.d. 0.88 Hz) indicating no mass binding (Figs 2C and S4). The ΔF difference between protein binding to the POPC/POPS and POPC SLBs was statistically significant (pvalue 6.94×10 −7 ). Together, this comparison showed that the binding of C protein to SLBs was dependent on the presence of charged POPS (Figs 2C and S4). Interestingly, the ΔD upon C protein adsorption to the POPC/POPS SLBs was very low (3.8×10 −7 , s.d 1.4×10 −6 ) indicating that the C protein forms a very rigid layer when adsorbed to the membrane [59]. This, in turn, indicates that the C protein does not bind as non-specific aggregates as these would have changed the viscoelastic properties of the system.
To probe the electrostatic nature of the initial C protein recruitment to the membrane, the C protein was injected in a buffer containing 1 M NaCl on SLBs preequilibrated with the same buffer. After reequilibrating the SLBs with TBS, the ΔF value was on average −0.6 Hz (s.d 0.3 Hz) compared to the pre 1 M NaCl equilibration showing that no C protein bound to the SLBs (Figs 2C and S4). The difference between injection on pure POPC SLBs and on POPC/POPS SLBs in 1M NaCl was not statistically significant (p-value 0.13). To test if the C protein membrane contacts are purely electrostatic, C protein was bound to SLBs in TBS, and then washed with 1 M NaCl buffer either right after equilibration or after a 2 h post-equilibration incubation. A wash with 1M NaCl resulted in what is referred to as "buffer shift" in the F and D values, which was reversed when moving back to the original buffer [60]. In both cases, only partial detachment of the C protein from the SLBs was detected with the final ΔF value being −11.9 Hz in the immediate wash experiment, and −11.3 in the 2h-incubation experiment (s.d. 1.9 and 1.5 Hz, respectively) ( Fig 2D). The ΔF values were significantly different from zero with p-values of 0.013 (immediate wash experiment) and 0.009 (2h-incubation experiment). However, they were not significantly different between the experiments with a p-value of 0.10. This shows that initial membrane recruitment of C protein is strongly dependent on its interactions with the negatively-charged lipid head groups. However, once bound, the C protein-

PLOS PATHOGENS
Membrane and RNA binding by tick-borne encephalitis virus capsid protein membrane interaction is complemented with non-electrostatic interactions such as membrane insertion or protein oligomerization within the bound layer. Since the amount of detached protein is not dependent on incubation time, it suggests that the protein´s non-electrostatic organization on the membrane occurs concurrently with adsorption.
When the same measurements were performed with the C 18-93 construct, similar results to the wild-type C were obtained. The C 18-93 protein injection caused a 24.1 Hz drop in ΔF (s.d 0.8 Hz) with the final ΔF value being significantly different from the C protein binding to pure POPC experiment (p-value 6.04×10 −7 ) (Figs 2C and S4). When the bound C 18-93 was washed with high-salt buffer, mass was retained on the SLB with a final ΔF value of −7.3 Hz (s.d. 0.2 Hz) (Figs 2D and S4). The final ΔF value was significantly different from zero (p-value 0.0002). Additionally, the final ΔF value was not significantly different from the full-length C protein experiments (p-values 0.079 and 0.066 for the comparison to the immediate wash and 2h incubation experiments respectively). This showed that the deletion of the N-terminus did not significantly alter membrane binding of the C-protein.

C protein inserts into membranes
To investigate if the C protein membrane binding includes insertion into the membrane after the initial electrostatic binding, we used Langmuir-Blodgett trough monolayer experiments. In this approach, the insertion of a protein into a lipid monolayer can be detected by following the pressure (π) of the monolayer after protein injection into the aqueous subphase, with increases in π corresponding to protein injection [60,61]. We formed monolayers of pure POPC or POPC/POPS mixture (60:40%, w/w) and monitored the π value of the monolayers following a C or C 18-93 injection into the subphase. The π increase (Δπ) between the starting (π 0 ) and ending pressures were measured at different π 0 values. The regression between Δπ and π 0 was used to determine the maximum injection pressure (MIP) of the proteins.
Injection of full-length C into the subphase led to sharp π increases at all tested π 0 values in monolayers containing 40% POPS (Figs 3A and S5). The MIP of full-length C was 39.1 mN/m, which is remarkably higher than the pressure of a lipid bilayer (~30 mN/m), suggesting that C is capable of inserting into bilayers such as the ER membrane ( Fig 3B) [62][63][64]. Contrastingly, in the absence of POPS, smaller π increases were detected, and the Δπ value was zero at π 0 of 20.20 mN/m (Figs 3A and S5). Furthermore, the MIP value of full-length C in pure POPC monolayers was 19.7 mN/m, which is notably smaller than the pressure of a lipid bilayer showing that negatively-charged lipids such as POPS are needed for C protein membrane insertion at physiological membrane pressures ( Fig 3C) [62][63][64]. The C 18-93 protein behaved similarly to the wild-type C in POPS-containing monolayers with a MIP of 37.7 mN/m (Figs 3A and S5). This corroborates the similar QCM-D results between C and C 18-93 , and shows that the N-terminus is not needed for C protein membrane insertion.

Membrane-bound C protein can recruit TBEV genomic RNA
Since RNA binding is crucial for the NC assembly, we investigated the C protein's ability to perform this function. First, we confirmed the binding in solution using in vitro transcribed TBEV genomic RNA with a gel electrophoretic mobility shift assay (GEMSA). Using a C to RNA molar ratio of 624:1 fully prevented the RNA from migrating normally indicating binding ( Fig 4A).
C protein presumably interacts simultaneously with membranes and RNA, and we therefore used QCM-D to study the C protein's RNA recruitment capability whilst membranebound. In these experiments, C protein was essential for the recruitment of TBEV genomic RNA on SLBs (Fig 4B-4D). Injecting 5 μg of in vitro transcribed TBEV genomic RNA on SLBs showing RNA binding (Fig 4B and 4C). In contrast, the ΔF value with SLBs with no prebound C remained close to zero after RNA injection, with an average equilibrium ΔF of −0.6 Hz (s.d. 0.3 Hz) indicating no binding of mass on the SLBs (Figs 4C and S6). The equilibrium ΔF value difference between C protein-containing and protein-free SLBs was statistically significant (p-value 1.34×10 −6 ) ( Fig 4C). Interestingly, the ΔD value also increased to 7.92×10 −6 on average after RNA binding (s.d. 1.03×10 −6 ) in the protein-treated SLBs, indicating that the RNA did not bind as a rigid uniform layer but rather as a flexible assembly (Fig 4B and 4C). Without prebound C protein, the ΔD did not increase and remained at an average −8.8×10 −8 at equilibrium (s.d. 5.4×10 −8 ) (Figs 4C and S6). The equilibrium ΔD value difference between the experiments with and without prebound C protein was statistically significant (p-value 0.008)

PLOS PATHOGENS
Membrane and RNA binding by tick-borne encephalitis virus capsid protein ( Fig 4D). These data show that membrane-bound C protein is capable of recruiting TBEV genomic RNA at the membrane, suggesting that this also happens in the context of NC assembly.
When the SLB RNA-binding experiments were repeated with C 18-93 , the results were similar as with full-length C. After RNA injection on C 18-93 -treated SLBs, the ΔF value dropped to −40.1 Hz on average and the ΔD value increased to an average of 5.7×10 −6 (s.d. 2.5 Hz and 2.1×10 −7 , respectively). The difference to the no protein experiment for both ΔF and ΔD was significant (p-values 0.002 and 0.0007, respectively) (Fig 4C and 4D). While the ΔF and ΔD values in the C 18-93 experiments were closer to zero than in the wild-type C experiments, the differences were not statistically significant (p-values 0.156 and 0.097 respectively).

TBEV virions contain negatively-charged phosphatidylserine
If the binding of C protein to negatively-charged lipid head groups is important for particle assembly, it is likely that these lipid species are incorporated into the virion during the budding into the ER. To investigate this possibility, we used UHPLC-QTOF-MS to characterize the TBEV lipid content of purified virions. The average integrated peak area of each investigated lipid species was compared between TBEV-containing samples, and buffer-only mock samples. Significantly higher integrated peak areas were interpreted as the presence of the species in the TBEV preparation, instead of noise caused e.g. by buffer or instrument contamination.
Multiple lipid species of different classes were detected in the TBEV samples ( Fig 5A and S1  Table). The virions contained lipids from two negatively-charged classes, phosphatidylserine (PS) and phosphatidylglycerol (PG) (Fig 5A and S1 Table). In addition, neutral carnitine fatty acids (Car), phosphatidylethanolamines (PE), PIs, phosphatidylcholines (PC), ceramides (Cer), hexosylceramides (HexCer), and triglycerides (TG) were detected (Fig 5A and S1  Table). The analysed lipid classes from the in-house database that were not found in virions were lyso-PE (LPE), lyso-PC (LPC), diglyceride (DG), and cardiolipin (CL) (Fig 5B and S1  Table). The virions contained multiple PS species: 32:1, 34:1, 36:1, 36:2, and 38:5 (number of carbons:number of double bonds) (Fig 5B and S1 Table). This confirms the biological relevance of using PS lipids in the model membrane experiments, especially since a total carbon to double bond ratio of 34:1 corresponds to POPS. As C protein was shown not to bind to PE, PI, and PC in the liposome co-sedimentation assays (see above), these lipids are likely incorporated into the virion without direct interaction with the C protein. Both of the negativelycharged lipids, PS and PG may be incorporated into the virion via direct interactions with the C protein, as we have shown that these interactions are charge, but not head group dependent in the liposome co-sedimentation assay. PS and PG interactions may be important for the coordination of NC assembly.

Discussion
Our unique combination of biophysical methods and model membrane systems allowed us to show that negatively-charged lipids are required for TBEV C protein membrane interactions, and that the protein inserts into membranes. The methods allowed us to demonstrate RNA recruitment onto membranes by the C protein, supporting that the NC assembly occurs on the ER membrane [7]. We show that residues 1-17 are not required for these interactions, and that membrane binding and RNA recruitment can occur independently of any other host or viral proteins. Additionally, we confirm the biological relevance of the interactions with negatively-charged lipids observed in the biophysical experiments by showing that negativelycharged PS and PG lipids are incorporated into the virion. PS has also been found previously

PLOS PATHOGENS
Membrane and RNA binding by tick-borne encephalitis virus capsid protein in the lipid composition of West Nile virus [47]. Our findings on C protein membrane binding properties agree with previous observations from TBEV and other flaviviruses [40][41][42][65][66][67].
N-terminally truncated flaviviral C proteins have been shown to be assembly competent and in vitro, able to bind RNA, which is consistent with our results with N-terminally truncated TBEV C protein [41,[68][69][70]. Instead, a role for the N-terminus may be in the reported modulation of host responses to infection [39,[71][72][73][74][75][76][77][78]. The membrane insertion directly detected in our experiments is central to C protein function. Other studies have found that deletions in the hydrophobic region of the α 2 helix significantly impair particle assembly  Table. https://doi.org/10.1371/journal.ppat.1011125.g005
While it is clear that the C protein can perform both membrane binding and insertion as well as RNA binding simultaneously, the mechanism for this remains unclear. Based on the QCM-D results, the membrane binding followed two different kinetics, and after binding and insertion, only half of the C protein could be removed by a high NaCl wash (Figs 2B and 4B). This suggests the formation of higher-order oligomers on the membrane. Most probably one set of dimers are recruited for instance by the α 4 -α 4 interface by charge interactions and insert into the membrane. Then they bind additional C protein molecules in another orientation, but they cannot insert due to steric hindrance which leaves their α 4 -α 4 interface or other charged surfaces exposed for RNA recruitment. Only dimers have been detected in TBEV virions by cross-linked mass spectrometry, but in order to form the NC, it is likely that higher-order C protein interactions are needed [37]. In cross-linking experiments with purified TBEV C protein, a population of trimers were detected which shows that higher-order oligomerisation is possible [34]. However, we cannot rule out that our C protein preparation has microheterogeneity despite its appearance in SEC, which could cause the partial insertion observed. Further functional studies are needed to determine the orientation, conformation and multimerisation status of C on the membrane.
Another factor that could influence the C protein interaction with negatively-charged lipids, would be its phosphorylation state. In West Nile virus, C protein, phosphorylated by protein kinase C, is imported into the nucleus through binding to importin α. The nuclear import leads to localization in the nucleolus where C binds to HDM2 starting apoptosis [79]. Phosphorylation decreases the cytoplasmic C protein amount, and reduces binding to RNA [80], thus decreasing the pool of protein available for NC assembly. TBEV C protein is also transported to the nucleolus, likely through phosphorylation-mediated binding to importin α as well [34,53]. Likely residues for phosphorylation are S19, S65 S83. Since S83 is present on the α 4 helix, phosphorylating it could repel the negatively-charged membrane and prevent binding, whereas dephosphorylation would promote assembly, increasing the propensity of C to bind to both membranes and RNA in TBEV. Such a mechanism for controlling the binding to negatively-charged lipids has recently been reported for herpesvirus nuclear egress complexes [81].
Recently, it has been shown that the neutralization of the C protein surface positive charge is important for RNA binding in the distantly-related Dengue virus (DENV) [82]. The recruitment of C protein to the membrane, that we demonstrate is dependent on negatively-charged head groups, provides a biologically relevant mechanism for charge neutralization on the C protein surface that interacts with the lipids. The remaining surface charge can be then neutralized by RNA recruitment. Therefore, negatively-charged ER lipid patches may act as the nucleation sites for NC assembly by binding the C protein that subsequently recruits the genomic RNA. PS, which we detected in the virion envelope, is also a major component of the ER [83]. ER lipid microdomains are most prevalent at the contact sites between the ER, lipid droplets, and mitochondria (rich in PG) [83,84]. These sites are associated with extensive exchange of lipids between organelles [83,84]. Mitochondria and lipid droplets are often observed in close proximity to the ER replication sites of TBEV, DENV and Zika virus [49,50,85]. Additionally, the recruitment of the C protein to lipid droplets has been suggested to be important for DENV assembly [51,86]. As flavivirus infections cause changes in the host cell lipid metabolism, it is possible that this leads to an increase in PS and PG content in the ER. The observed frequent association of mitochondria and lipid droplets with flavivirus replication sites as well as the presence of PS and PG in the TBEV envelope maybe the results of this [46,47,86].
In conclusion, we have determined that NC assembly is most likely membrane-associated, involving specific lipids. Although often ignored in virus structure and assembly, the importance of specific lipid species is gradually being recognised in flavivirus assembly, including the recent discoveries of ordered lipids in flavivirus membranes coordinated by the E and M membrane helices [6,[87][88][89][90]. Our results provide a starting point for further structural and mechanistic studies on the role of lipid interactions in TBEV NC assembly.

Cloning and protein purification
Full-length TBEV C protein (residues 1-96) and a truncated C protein (C 18-93 , containing residues 18-93) from the Kuutsalo-14 strain (GenBank: MG589938.1) with an N-terminal 6x His + Small Ubiquitin-like Modifier (SUMO) tag were cloned into pET28 plasmid as a commercial service by DNA Dream Lab (Helsinki, Finland) [57]. The proteins were expressed in E. coli ArcticExpress cells. The cells were grown o/n at 37˚C with 225 RPM shaking in lysogeny broth (LB, 1% w/v tryptone, 0.5% w/v yeast extract, 85 mM NaCl) containing 50 μg/ml kanamycin and 10 μg/ml gentamycin. Cells were inoculated into fresh LB without antibiotics, and grown at 30˚C with 225 RPM shaking until OD600 reached 0.5. The cultures were moved to +12˚C with 225 RPM shaking for an hour and induced with isopropyl β-d-1-thiogalactopyranoside (IPTG, final concentration 1 mM). The induced cultures were grown for 24 h at 12˚C with 225 RPM shaking and harvested by centrifuging at 3000 × g for 20 min at +4˚C. The supernatant was discarded, and the pellets stored at −80˚C until needed.
The frozen pellets were resuspended on ice in 300 mM NaCl, 50 mM Tris, 375 mM L-arginine, 20 mM imidazole, pH 7.5 containing Pierce protease inhibitors (Thermo Fisher, A32965). The cells were lysed with an Emulsiflex apparatus (Avestin Inc., Ontario, Canada) operated at 15 000 psi for 10 min. The cell lysates were precleared for 30 min at 20 200 × g at + 4˚C. The resulting supernatant was further cleared by centrifuging at 96 600 ×g for 1.5 h at +4˚C. The proteins were captured with nickel immobilized metal affinity chromatography using a 5 ml HisTrap FF column (Cytiva, Massachusets, USA) and a linear imidazole gradient from 0 mM to 1000 mM over 25 column volumes. The purified proteins were bufferexchanged into 300 mM NaCl, 50 mM Tris, pH 7.5 with PD 10 columns (Merck, New Jersey, USA) using the manufacturer's protocol. The SUMO tags were cleaved using in-house-produced SUMO protease o/n (full-length C) or for 48 h (C 18-93 ) at room temperature (RT) with gentle agitation. The cleaved tags were removed by cation exchange chromatography using a 1 ml HiTrap SP HP column (Cytiva) with a linear NaCl gradient from 300 mM to 1000 mM over 25 column volumes. SEC was used to buffer exchange the samples into 300 mM NaCl, 50 mM Tris, pH 7.5 using a Superdex 200 10/300 GL column (Cytiva). The peak fractions were collected, aliquoted, and stored in −80˚C. The aliquots were characterized with SDS-PAGE and immunoblotting (see below), the protein concentrations were measured with a QFX fluorometer (DeNovix, Delaware, USA) using a Qubit Flex kit (Thermo Fisher, Massachusetts,

PLOS PATHOGENS
Membrane and RNA binding by tick-borne encephalitis virus capsid protein USA), and the A 260 /A 280 ratios of the samples were measured using a Nanodrop instrument (Thermo Fisher).

SDS-PAGE and immunoblotting
Samples were mixed with Laemmli sample buffer, incubated for 5 min at 95˚C, and proteins were resolved using electrophoresis in 4-20% polyacrylamide gels (Bio-Rad, California, USA). Proteins were visualized using Coomassie blue staining or by immunoblotting. For immunoblotting, the proteins were transferred onto nitrocellulose membranes (GE Healthcare, Illinois, USA) using a Trans Blot turbo apparatus (BioRad) using the manufacturer's protocol. The membranes were blocked using 5% (w/v) milk in 150 mM NaCl, 20 mM Tris, 0.1% TWEEN 20, pH 7.6 (TBST) for 30 min at RT with gentle rocking and washed with TBST. Membranes were incubated with the primary anti-C antibody (1:1000 dilution in TBST containing 5% milk) [52]. After incubation, the membranes were washed 3 times with TBST for five minutes with gentle rocking. The secondary antibody was diluted 1:10 000 in TBST, added on the membranes, and the membranes were incubated for 30 minutes at RT with gentle rocking. The secondary antibody used was IRDye 680 RD goat anti-rabbit IgG (Li-COR). After incubation, the membranes were washed with TBST and visualised using an Odyssey imager (Li-COR).

Liposome co-sedimentation assay
The protocol was adapted from reference [95]. Liposomes of various lipid species (see Table 1) were prepared by diluting total of 0.5 mg of lipids in a 10:3 mixture of chloroform and methanol in glass tubes. Lipid films were prepared by drying the lipids under a gentle nitrogen

PLOS PATHOGENS
Membrane and RNA binding by tick-borne encephalitis virus capsid protein stream while rotating the glass tubes. The lipids were further dried under a gentle nitrogen stream for 30 min. The lipids were rehydrated to 1 mg/ml by adding 500 μl 150 mM NaCl, 50 mM Tris, pH 7.6 (TBS) and by incubating at RT for 30 min. The liposomes were sonicated with a bath sonicator for 20 s (Transsonic T310, Elma Schmidbauer, Singen, Germany) and either used immediately, or stored at +4˚C for up to two days. Fresh aliquots of full-length C protein were thawed on ice and centrifuged for 10 min at 20 000 × g at + 4˚C, and the supernatant collected. The protein was incubated either with experimental liposomes or TBS for 5 min at RT with a final protein concentration of 7 μM and a final lipid concentration of 500 μg/ml. The liposomes were pelleted by centrifuging for 20 min at 100 000 × g at RT. The supernatants were collected, and the pellets resuspended in 50 μl TBS. Both the supernatants and the pellets were analysed with SDS-PAGE as above, and the fraction of C protein co-sedimenting with the liposomes was quantified using Fiji [96].

Liposome flotation assay
The protocol was adapted from reference [97]. Liposomes containing either DOPC/DOPE/r-DHPE/POPS (15:14:1:40%, w/w) or DOPC/DOPE/r-DHPE (60:39:1%, w/w) were prepared as for the co-sedimentation assay, except that they were rehydrated in TBS containing 0.3 M sucrose. Purified full-length C protein was cleared as above and incubated with TBS containing 0.3 M sucrose or the liposomes (protein concentration 7 uM, lipid concentration 415 mg/ml) for 10 min at RT in 1 ml ultracentrifuge tubes. The step gradient was created by adding 60% sucrose in TBS (w/w) to yield a final sucrose concentration of 40% (w/w) to the protein-liposome mixture at the bottom of the tube. This was overlaid first with 25% (w/w) sucrose in TBS, followed by a layer of TBS. The samples were centrifuged for 30 min at 259 000 × g at + 4˚C. Samples (100 μl aliquots) were collected from the TBS phase (fraction 1), the liposome-containing TBS-25% sucrose phase boundary (fraction 2), the 25% (w/w) sucrose-containing phase (fraction 3), and the 40% (w/w) sucrose-containing phase (fraction 4). SDS was added to the samples to yield a final concentration of 0.1% (w/v), and the samples were mixed at RT. The relative protein content of the fractions was quantitated with SDS-PAGE and Fiji (see above). The relative lipid content of the fractions was quantitated by diluting the samples 1:50 in TBS and measuring the r-DHPE fluorescence using a VICTOR Nivo plate reader (PerkinElmer, Massachusetts, USA) with excitation set at 570 nm and emission set at 595 nm.

RNA synthesis and electrophoretic mobility shift assay
Full-length TBEV genomic RNA was produced from an infectious clone based on the TBEV 93/783 sequence (GenBank: MT581212) using in vitro transcription. Linearized DNA containing the clone was transcribed using the mMESSAGE mMACHINE SP6 Transcription kit (Thermo Fisher) according to the manufacturer's protocol. RNA was purified by LiCl precipitation according to the transcription kit protocol. RNA concentration was measured using a nanophotometer (DeNovix).
Transcribed RNA (300 ng) was mixed with C protein at molar ratios ranging from 1:2 to 1:2496 in 150 mM NaCl, 50 mM Tris, pH 7.5 prepared with RNAse-free reagents (RNA buffer) and incubated on ice for 20 min. The samples were mixed with loading dye (final composition: 60% v/v formamide, 0.0133% w/v bromophenol blue, 0.0053% v/v xylene cyanol, 0.72% w/v Tris, 0.37% w/v boric acid, 6.6 mM EDTA) and analysed with electrophoresis on a 0.5% agarose gel containing SYBR Green I (Thermo Fisher) for 50 min at 70 V.

SLBs and QCM-D
The protocol was adapted from reference [60]. Liposomes containing either pure POPC, or a mixture of POPC and POPS (60:40% w/w) were prepared as above in 20 mM citric acid, 50 mM KCl, 0.1 mM EDTA, pH 4.5 (CKB). The liposomes were homogenized by extruding 11 times through a 50 nm polycarbonate filter (Nuclepore Track-Etched Membranes, Whatman, Maidstone, UK) using a Mini Extruder instrument (Avanti Polar Lipids). QCM-D experiments were conducted with an X4 instrument equipped with a flow chamber (AWSensors, Valencia, Spain) using wrapped 14 mm 5 MHz, Cr/Au-SiO 2 polished QCM-D sensors (AWSensors) at +23˚C. The sensors were incubated overnight in 2% SDS and cleaned with a UV-ozone cleaner (Bioforce Nanosciences, Iowa, USA) for 30 min before use. Frequency and dissipation were monitored for overtone 3, although they were measured for overtones 1, 5, 7, 9, and 11 as well. The chambers were filled with CKB, and the SLBs were formed by injecting 150 μl of the liposomes at a concentration of 0.1 mg/ml. After the formation of the SLBs, TBS was injected until the system reached equilibrium. For the 100% POPC SLBs, 7.7 μg of C protein in 100 μl TBS (corresponding to a concentration of 7 μM) was injected. For the 60% POPC/40% POPS SLBs, full-length C or C 18-93 was injected either as above or the SLBs were washed with 1000 mM NaCl, 50 mM Tris, pH 7.6 (HSQB) and C protein injected as above but in HSQB. After C protein injection, the chambers were washed with TBS until equilibrium. In the experiments where C protein was injected in TBS, the protein-containing SLBs were washed with HSQB either immediately after reaching equilibrium, or after 2h incubation. Similarly, in C 18-93 experiments, the protein-containing SLBs were washed with HSQB immediately after equilibrium was reached. After HSQ wash, the SLBs were washed with TBS until equilibrium. For the RNA experiments, the SLBs were washed with RNA buffer either without protein addition, or after an injection of C or C 18-93 as above. 5 μg of in vitro transcribed TBEV genomic RNA in 100 μl RNA buffer was injected, followed with a wash with RNA buffer, and a wash with TBS until equilibrium. The equilibrium ΔF and ΔD values for each step were calculated as the average of a 30 s period starting 400 s after the start of the final TBS wash.

Lipid monolayer experiments
The protocol was adapted from reference [60]. Lipid formulations of either pure POPC or a 60:40% (w/w) mixture of POPC and POPS were prepared in chloroform (final lipid concentration 0.5 mg / ml). A Microtrough G1 system (Ø 53 mm x 4 mm, Kibron, Helsinki, Finland) was used for the experiments. The trough was covered to prevent evaporation, and the experiments were conducted at +21˚C. The subphase was formed of TBS and constantly stirred with a magnetic stirrer during the measurement. π was measured with a Wilhelmy paper plate presoaked in TBS and calibrated for each measurement. The monolayers were formed by adding the lipid solutions on to the subphase in a dropwise manner and letting the chloroform evaporate. Once the initial π 0 value of the experiment was reached, the monolayer was left to equilibrate for 30 min. Then, C protein or C 18-93 in TBS was injected with a syringe under the monolayer via an injection port to reach the final protein concentration of 30 nM in the subphase. The π value was monitored until it plateaued. The π 0 values were calculated as an average of the 30 s preceding the injection, and the final π values were calculated as the average of the 30 s surrounding the highest value reached. The MIP-values were determined by plotting Δπ against π 0 and extrapolating the Δπ/π 0 plot to the x axis.

Cell culture
Human neuroblastoma SK-N-SH cells (gift from Prof. Olli Vapalahti, University of Helsinki) were maintained in low-glucose Dulbecco 0 s Modified Eagle 0 s Medium (DMEM, Sigma)

Virus propagation and titration
TBEV-Eu strain Kuutsalo-14 passage 1 (gift from Prof. Vapalahti) was used to infect confluent SK-N-SH cells using a multiplicity of infection of 0.003 [57]. The cells were washed twice with phosphate-buffered saline (PBS), virus was added in infection medium (low-glucose DMEM, 0.5 mg/ml penicillin, 500 U/ml streptomycin, 2 mM glutamine, 2% FBS, 35 nM rapamycin [98]), and incubated at +37˚C in 5% CO 2 . At 72 h post-infection, the supernatant was collected, and centrifuged for 10 min at +4˚C at 3800 × g. After centrifugation, the supernatant was collected and immediately purified (see below), or titered and stored at -80˚C.
For titration, virus samples were serially diluted 10 fold in infection medium. Confluent SK-N-SH cells in 6-well plates were washed twice with PBS, and 200 μl of the virus dilutions or infection medium was added on the cells. The cells were incubated at +37˚C for 1 h with gentle shaking every 5 min. After the incubation, 3 ml of minimum essential medium (MEM, Gibco) containing 0.5 mg/ml penicillin, 500 U/ml streptomycin, 2 mM glutamine, 2% FBS, and 1.2% Avicell was added to each well. The cells were incubated at +37˚C in a 5% CO 2 atmosphere for 4 days, after which the medium was removed, the cells washed with PBS, and fixed with 10% formaldehyde for 30 min at RT. The plaques were visualised by incubating the fixed cells with 0.5% crystal violet for 10 min and washing with water. The viral titre was expressed as plaqueforming units per ml (pfu/ml).

Virus purification
The virus was precipitated from the cell supernatant by adding 8% (w/v) polyethylene glycol (PEG) 8000 (Sigma-Aldrich), incubating at +4˚C with gentle mixing for 3 h, and pelleted at +4˚C, at 10,500× g for 50 min. The supernatant was discarded, and the pellets were carefully washed with 20 mM HEPES, 150 mM NaCl, 1 mM EDTA, pH 8.5 (HNE) and resuspended by incubating in HNE overnight at +4˚C on an orbital shaker.
The dissolved pellets were loaded onto linear 10-70% sucrose gradients prepared in HNE. The samples were centrifuged at +4˚C at 44,500 × g for 17 h and the virus-containing lightscattering bands were collected. The samples were buffer exchanged to HNE using Slide-A-Lyzer dialysis cassettes (MWCO 10 kDa; ThermoFisher) at +4˚C for 1h with gentle agitation and the HNE was replaced with fresh HNE, and the dialysis continued overnight. The titre was determined using plaque titration.

Mass spectrometry analysis of TBEV lipids
Purified virus (7×10 8 pfu) in HNE or just HNE was mixed with methanol and chloroform (1:3:1 ratio of sample:methanol:chloroform), and vortexed for 30 s. Insoluble components were precipitated by centrifuging for 10 min at +4˚C at 20 000 × g and the supernatants were stored in glass vials with a nitrogen atmosphere in -80˚C.
The data were processed using Batch Targeted Feature Extraction algorithm within Mas-sHunter™ ProFinder version B.10.02 (Agilent Technologies Inc., Santa Clara, CA, USA). An in-house database with exact mass and experimental retention times of lipids was used for identification.

Statistical procedures
Statistical significance of the results were assessed either using one sample T-test (assessment of remaining F signal in 1M NaCl wash QCM-D experiments) or Welch's T-test (all other experiments) performed in R 4.2.1.