Glucagon-like peptide-2 receptor is a receptor for tick-borne encephalitis virus to infect nerve cells

Tick-borne encephalitis virus (TBEV) is a tick-borne avivirus that causes severe encephalitis disease1,2. Host proteins required for TBEV entry remain largely unknown3. Here we performed a genome-wide CRISPR-Cas9 knockout screen and identied G-protein-coupled receptor glucagon-like peptide-2 receptor (GLP2R) as a receptor for TBEV to infect nerve cells. Knockdown or knockout of GLP2R reduced TBEV infection of different nerve cells; trans supply of GLP2R restored viral infection. GLP2R directly binds to viral envelope domain III through its extracellular loop 1 (ECL1). TBEV infection can be blocked by the ECL1 peptide, a functional ligand to GLP2R, or GLP2R antibodies. GLP2R-decient mice were generated to validate the role of GLP2R in TBEV infection and pathogenesis. Wild-type mice succumbed to TBEV infection and developed >107 TCID50 (median tissue culture infectious dose) virus per gram of brain tissue. In contrast, all GLP2R-decient mice survived TBEV infection without detectable infectious virus in brain. Altogether, our results support GLP2R as a receptor for TBEV to infect nerve cells.

ECL1 peptide, a functional ligand to GLP2R, or GLP2R antibodies. GLP2R-de cient mice were generated to validate the role of GLP2R in TBEV infection and pathogenesis. Wild-type mice succumbed to TBEV infection and developed >107 TCID50 (median tissue culture infectious dose) virus per gram of brain tissue. In contrast, all GLP2R-de cient mice survived TBEV infection without detectable infectious virus in brain. Altogether, our results support GLP2R as a receptor for TBEV to infect nerve cells.

Main Text
Host factors are essential for viral replication and pathogenesis. To identify host proteins critical for TBEV infection cycles, we performed a genome-wide screen using a lentivirus based CRISPR-Cas9 system 4,5 . The CRISPR-Cas9 system delivered single-guide RNAs (sgRNAs) targeting 19,050 human genes. Human glioma T98G cells were transduced with pooled lentivirus and challenged with TBEV Far Eastern subtype (TBEV-FE, strain WH2012) 6  Glucagon-like peptide-2 receptor (GLP2R) emerged as a top candidate from the screen (Extended Data Fig   2a & e). GLP2R belongs to G protein-coupled receptor (GPCR) superfamily B 8,9 . It is mainly expressed in the central nervous system and enteric neurons [10][11][12] . To examine its function in TBEV infection, we performed a GLP2R knockdown experiment in two human glioma cells (T98G and U251) (Extended Data   Fig 3a & b). Knockdown of GLP2R signi cantly impaired the infection of both TBEV-FE and TBEV European subtype (TBEV-Eu, strain Neudoer ) (Extended Data Fig 3c-f), as evidenced by >23-fold reduction in infectious virus yields (Extended Data Fig 3f). To con rm the knockdown results, we knocked out GLP2R (glp2r -/-) in T98G and U251 cells using CRISPR-Cas9 with multiple sgRNAs (Supplementary  Table 3 and Extended Data Fig 4a-c). The knockout of GLP2R did not affect cell viability, but signi cantly reduced viral replication of TBEV-FE and TBEV-Eu (Fig 1a & b and Extended Data Fig 5a-e). At 48 h postinfection, the glp2r -/cells produced >56-fold less infectious viruses than the parental normal cells (Fig. 1c & d). Trans expression of GLP2R in the glp2r -/cells restored the infectious virus production (Fig 1c & d). In contrast, the knockout of GLP2R did not affect the infection of other neurotropic viruses, including Japanese encephalitis virus (JEV), Zika virus (ZIKV), vesicular stomatitis virus (VSV), human cytomegalovirus (HCMV) and herpes simplex virus (HSV) (Fig 1e). Collectively, the results indicate that GLP2R is speci cally required for TBEV infection.
We validated the role of GLP2R in TBEV infection in other nerve cells, including human neuroblastoma SK-N-SH, human astrocytoma CCF-STTG1, and human primary astrocytes. Although these cells expressed different levels of GLP2R (Extended Data Fig 6a), knockdown of GLP2R decreased TBEV infection (Extended Data Fig 6b & c), con rming the function of GLP2R in TBEV infection in these nerve cells. Since TBEV can also infect non-human, non-neurogenic BHK-21 and Vero-E6 cells [13][14][15][16][17][18][19] , we examined GLP2R expression in these cells. No and low expression of GLP2R was found in BHK-21 and Vero-E6 (Extended Data Fig 6d); consequently, treatment with GLP2R siRNAs did not affect TBEV infection of these cells (Extended Data Fig 6e); however, ectopic expression of GLP2R enhanced TBEV infection of Vero-E6 cells. As a control, TBEV infection was also enhanced when glioblastoma U251 cells (with low endogenous GLP2R) expressed exogenous GLP2R (Extended Data Fig 6f). The results imply that GLP2R is speci cally required for TBEV infection of human nerve cells.
Since GLP2R is a cell membrane protein 20 , we examined its role in TBEV entry using a lentivirus-based pseudovirus system. Knockout of GLP2R signi cantly reduced the transducing e ciency of TBEV envelope-pseudovirus, but not the VSV glycoprotein-pseudovirus, on T98G cells (Fig 2a). Knockout of GLP2R also reduced the binding of authentic TBEV to cells at 4 °C (which allows virus binding but not entry), whereas over expression of GLP2R increased TBEV binding (Fig 2b). To analyze TBEV internalization, we quanti ed intracellular viral RNA after incubating the cells with TBEV at 37 °C infection for 1 h. Compared with normal T98G cells, less viral RNA was detected in the glp2r -/cells, whereas more viral RNA was detected in the GLP2R-overexpressing cells (Fig 2c). Consistent with the role of GLP2R in TBEV entry, pre-treatment of T98G cells with an anti-GLP2R mAb or a TBEV envelope mAb (A4) 21,22 suppressed viral infection (Fig 2d). Furthermore, incubation of T98G cells with H33D peptide 10,23 , a functional ligand to GLP2R, inhibited TBEV infection (Fig 2e & f). In contrast, incubation with D31D peptide, a non-functional H33D mutant that lacks the N-terminal two amino acids 23 , did not affect TBEV infection (Fig 2g). The results demonstrate that GLP2R contributes to TBEV attachment and entry.
To determine whether GLP2R affects viral translation and RNA replication, we transfected a luciferase reporter subgenomic replicon RNA of TBEV into control and glp2r -/-T98G cells. No signi cant difference in luciferase activity was detected between the two transfected cell types (Fig 2h), suggesting that GLP2R does not affect viral translation and RNA synthesis.
We mapped the domains required for the GLP2R and TBEV envelope interaction. TBEV envelope could e ciently pull-down human (huGLP2R) and less e ciently mouse GLP2R (moGLP2R) (Fig 3a). Pulldown experiments with a panel of deletion GLP2R and TBEV envelope (Extended Data Fig 7a &  A synthetic peptide Y41L, representing the ECL1 amino acid sequence, e ciently competed away the GLP2R/viral envelope interaction (Fig 3d). The dissociation equilibrium constants (K D ) between peptide Y41L and domain III of viral envelope (G99I-lys-biotin) or puri ed TBEV prM-E (premembrane-envelope) protein were estimated to be 172 nM or 76 nM, respectively (Fig 3e & Fig 8c). To further demonstrate the importance of ECL1 in viral envelope binding, we replaced different lengths of ECL1 by a linker Gly-Ser-Gly (GSG) in the context of GLP2R (Extended Data Fig 9a). These ECL1 mutant GLP2R variants remained expressed on the cytoplasmic membrane (Extended Data Fig 9b), but attenuated their binding to viral envelope by 60-66% (Extended Data Fig 9c). Consistently, the ECL1 mutant GLP2R (mutGLP2R=239GSG254) lost its ability to rescue TBEV infection of glp2r -/-T98G cells (Fig 3g). Besides ECL1, we also found the N-terminal extracellular domain (ECD) of GLP2R may modestly facilitate viral envelope binding (Extended Data Fig 9d); however, the ECD deletion GLP2R (GLP2R-ΔECD) could still rescue TBEV infection of glp2r -/-T98G cells (Fig 3g). Thus, the ECL1 of GLP2R is the key domain for interacting with TBEV envelop.

f, Extended Data Fig 8a & b). Functionally, the Y41L peptide inhibited TBEV infection of T98G cells (Extended Data
To evaluate the signi cance of GLP2R in TBEV infection in vivo, we generated Cas9-KO mice with a deletion of exons 2-10 in glp2r gene (Extended Data Fig 10a). The glp2r -/mice did not develop observable defects. After infected with TBEV-FE (3.2×10 7 PFU) via an intraperitoneal route, wild-type mice developed signi cant weight loss (Extended Data Fig 10b); 58.8% (n=10/17) of them succumbed to the infection (Fig 4a); all animals developed viremia (Fig 4b) and high viral loads in brain, spleen, and intestine (Fig 4c & d). Remarkably, none of the glp2r -/mice (n=13) developed weight loss (Extended Data Fig 10b); all survived the infection (Fig 4a); no infectious virus was detected from blood, brain, spleen, or intestine (Fig 4b-d). Only low levels of viral RNA (<10 5 RNA copies per g of tissue) were detected from the glp2r -/brain, spleen, and intestine on day 7 post-infection (Fig 4c). Since TBEV is an encephalitis virus, it is important to note that, on day 7 post-infection, the wild-type mice developed 2.4×10 7 TCID 50 infectious virus per gram of brain tissue, whereas no infectious virus was detected in the glp2r -/brain (Fig. 4c). The results demonstrate that GLP2R is essential for TBEV to productively infect mice.
We compared the histopathology of infected wild-type and glp2r -/mice. H&E staining showed multifocal lesions to the brain, spleen, and small intestine from the wild-type mice (Fig. 4e). No such histopathology was observed in the glp2r -/mice in general (Fig 4e) with only one mouse exhibiting mild pathological changes (data not shown). These ndings were supported by quantitative pathological scoring (Extended Data Fig 10c). Consistent with the in ammatory cell in ltration (Fig. 4e), RNAseq analysis showed signi cantly higher levels of chemokines in the brain of infected wild-type mice (lost >20% body weight) than those from the infected glp2r -/mice (Fig 4f). The histopathological results support the critical role of GLP2R in TBEV infection and pathogenesis.
Heparan sulfate proteoglycans (HSPGs) facilitate TBEV infection via interaction with viral envelope protein, allowing virion to engage in receptor binding 24 (Fig 4g). Our results have identi ed GLP2R as a host protein to mediate TBEV attachment and entry to neurogenic cells. In cell culture, TBEV infection is diminished when GLP2R expression is knocked down or knocked out, but the infection is rescued when GLP2R is trans complemented. In a mouse model, knockout of glp2r abolished viral replication and pathogenesis, leading to no morbidity and mortality. Biochemically, the viral envelop/receptor interaction is primarily mediated by an extracellular loop ECL1 of GLP2R and domain III of viral envelope. These in vitro and in vivo evidence strongly support GLP2R as a receptor for TBEV to infect neurogenic cells. However, since GLP2R-negative non-neurogenic cells are also susceptible to TBEV infection, we speculate that an unidenti ed alternative host factor can mediate TBEV entry to these cells. Nevertheless, it is important to point out that, due to the neurotrophic nature of TBEV, GLP2R, which is predominantly expressed in the central nervous system and enteric neurons [10][11][12] , could serve as a major receptor for TBEV infection in vivo. Our study has advanced the fundamental knowledge of avivirus biology and provided a new antiviral approach through blocking the viral envelope/GLP2R interaction.  Genome-Scale CRISPR-Cas9 Knockout Screen and data analysis. A pooled GeCKO v2 library encompassing 122,411 single-guide RNAs (sgRNAs) against 19,050 human genes derived by the Zhang laboratory 27 was obtained from a commercial source (Addgene, #1000000048). The lentivirus packaging, puri cation, titers using the GeCKO library were performed as previously described 4  Cell viability assay. Cell viability was evaluated using a Cell Titer-Glo Luminescent Cell Viability Assay kit (Promega) according to the manufacturer's instructions. In brief, 1×10 4 cells in 100 μl culture medium were seeded into opaque-walled 96-well plates for 24 h and 100 μl of Cell Titer-Glo reagent was added to each well. After a 5 min shaking and 10 min incubation, luminescence was measured by GloMax 20/20 (TurnerBio Systems).

Methods
GLP2R knockout cell lines. The GLP2R knockout cell lines were generated based on CRISPR/Cas9. The sequence of sgRNAs target to GLP2R are listed in Supplementary Table 3. The sgRNA were cloned into the lentiCRISPRv2 plasmids (Addgene, #98290) and packaged in HEK-293T cells by cotransfection with psPAX2 (Addgene, #12260) and pMD2.G (Addgene, #12259). To generate stable knockout cell lines, T98G and U251 cells at ~50% con uency were transduced with the sgRNA containg lentivirus at a MOI of 0.3 and selected in with 2 μg ml -1 puromycin for 10 days. The knockout e ciency was analyzed using western blotting assays and sanger DNA sequencing.
Expression construct. The wild type human or mouse GLP2R and its mutants with HA-tag or egfp-tag at the C-terminus were cloned into the pcDNA3.1(-) or pLVX-puro expression vectors. Flag-tagged TBEV envelope and its mutants were cloned into the pcDNA3.1(-) expression vector. The plasmids transfection was performed using Lipofectamine 3000 (Invitrogen) transfection reagent according to the manufacturer's protocol. A list of cloning primers is displayed in Supplementary Table 1. qRT-PCR. Total cellular RNA was isolated with TRIzol (Invitrogen) reagent according to the manufacturer's protocols. Viral RNA in culture supernatants was extracted with QIAamp Viral RNA Mini Kit (Qiagen) according to the manufacturer's instructions. The quanti cation of speci c gene transcripts was analyzed by one-step real-time qRT-PCR with speci c primers and the HiScript II One Step qRT-PCR SYBR Green Kit (Vazyme) on the Applied Biosystems QuantStudio 6 Flex. The primer sequences for qPCR/RT-PCR were designed using Primer3 software (see Supplementary Table 4). The data were normalized to levels of β-actin mRNA in each individual sample. The relative expression level and absolute quanti cation were calculated by 2 −ΔΔCt method and standard curve line respectively.
Western blotting (WB) and Co-immunoprecipitation (Co-IP). Whole-cell lysates for both WB and Co-IP were prepared using a lysis buffer containing 50 mM Tris-base (pH 7.5), 1 mM EGTA, 1 mM EDTA, 1% Triton X-100, 150 mM NaCl, 100 µM phenylmethylsulfonyl uoride (PMSF) and protease inhibitors (Roche) for 30 min at 4°C. Cell lysates were centrifuged at 14,000×g for 10 min at 4 °C and quanti ed using the Bradford method. For WB, the supernatants were recovered and denaturized at 95 °C for 10 minutes. For Co-IP, the supernatants were collected and mixed with Protein G-agarose (Millipore) and various antibodies for 16 h at 4 °C. After six washes using ice-cold lysis buffer, protein G agarose-bound immune complexes were then eluted and subjected to WB analysis. 30 μg of total protein from each sample was resolved by SDS-PAGE and transferred to nitrocellulose. Membranes were blocked with TBST (pH 7.4, containing 0.1% Tween-20) containing 5% skimmed milk for 1 h at room temperature, followed by incubation with anti-sera containing primary antibodies overnight at 4 °C (see Supplementary Table 6). Membranes were washed and incubated for 1 h at room temperature with the HRP-conjugated secondary antibodies. Membranes were imaged using the FluorChem HD2 system (Alpha Innotech). Images were analyzed using AlphaEaseFC software (Alpha Innotech).
Pseudovirus production and transduction. Pseudoviruses were produced by co-transfection 293T cells with psPAX2, pLenti-GFP-luc, and plasmids encoding either TBEV envelope or VSV-G by Lipofectamine 2000 (Invitrogen). The supernatants were harvested at 48 h post transfection, centrifuged and then passed through 0.45 μm filter. For transduction, cell in 24-well-plate were inoculated with pseudovirus containing medium overnight. Cells were lysed with passive lysis buffer (Promega) and luciferase activity were measured to calculate the transduction e ciency at 40 hours post transduction.
Virus binding and internalization assays. Control or glp2r -/cells were washed twice with ice-cold PBS and incubated with TBEV at an MOI of 20 in cold medium supplemented with 2% FBS on ice for 1 h. For the binding assay, the supernatant was removed, and cells were washed with ice-cold PBS six times. After washing, cells were collected, and RNA was measured by qRT-PCR. For the internalization assay, the washed cells were then incubated in medium supplemented with 2% FBS and 15 mM NH 4 Cl at 37 °C for 1 h. Cells were chilled on ice and treated with 500 ng/ml proteinase K on ice for 1 h. After three additional washes, cells were collected, and RNA was measured.
Blocking assays with antibodies and peptide. T98G cells were preincubated with isotype control or GLP2R antibodies at different concentration for 1 h, or serially diluted H33D, D31D and Y41L peptides for 2 h at 37 °C. After three washes, TBEV (MOI of 0.5) were added and incubated for 24 h. Cells were collected, and viral RNA was measured by qRT-PCR. The peptides used for blocking assays in this study were purchased from GLSBioChem and listed in Supplementary Table 7.
Puri cation of prM-E recombinant fusion protein. The DNA sequence encoding prM (residues 1 to 163), sE (residues 1 to 401), a N-terminal signal peptide (MGILPSPGMPALLSLVSLLSVLLMGCVA) for secretion and a His 10-tag at the C-terminus was ampli ed by standard PCR techniques. The transmembrane region of prM (residue 126 to 163) was replaced by a linker containing a Tobacco Etch Virus (TEV) protease recognition site (GENLYFQG). The furin cleavage site of prM (85-RTRR-88) was mutated to (85-TTRS-88) to prevent cleavage of the recombinant protein by intracellular proteases.
The prM-E recombinant fusion protein was expressed in pExpi293 cells (Thermo Fisher Scienti c) at 33°C for 5 days and puri ed on a 5-ml Talon Cobalt column. After sample application, the column was washed with 50 ml of buffer A (25 mM phosphate pH 8.0, 300 mM NaCl, 7 mM imidazole), and protein was eluted with a 100-ml linear gradient to 100% buffer B (25 mM phosphate pH 8.0, 300 mM NaCl, 500 mM imidazole). Fractions containing prM-E protein were pooled, concentrated and buffer exchanged into PBS in a 30-kDa MWCO spin concentrator. Proteins were ash-frozen in liquid nitrogen and stored at −80°C .
Octet binding assay. Octet binding experiments were performed on an Octet RED96 system (ForteBio, Fremont, CA). The puri ed biotinylated TBEV prM-E protein or domain III of viral envelope (E DIII, G99I-lysbiotin) was immobilized on a Streptavidin Biosensor (ForteBio) at 50 μg mL −1 in PBS with 0.02% Tween 20 and 0.1% BSA. GLP2R ECL1 peptide (Y41L) was diluted to different concentrations with PBS with 0.02% Tween 20 and 0.1% BSA. 1200s and 600s association/dissociation processes were performed for Y41L binding to E DIII and prM-E respectively. The resulting data were background subtracted and from the dissociation and association curves, k a , k d and K D (k d /k a ) were calculated using Octet data acquisition 6.4, ForteBio data analysis software 6.4 (Refer to the data in the Extended Data Fig 8b). Additionally, the equilibrium responses were estimated from the association curves using a 2-phase exponential association equation, these responses were plotted against Y41L concentrations to obtain equilibrium binding constant K D in GraphPad Prism 6 using a single site binding model (Refer to the data in the Fig 3e&f).
Mouse experiments. The animal experiments were approved by the Animal Care and Use Committee at the Wuhan Institute of Virology, and conducted in conformity with the Guide for the Care and Use of Laboratory Animals of the Wuhan Institute of Virology in ABSL3 facility (Ethic number WIVA02202002).
glp2r −/− mice purchased from Jiangsu GemPharmatech Co., Ltd, were generated in the C57BL/6J background by depleting exons 2~10 of the glp2r using CRISPR/Cas9 technology (see Supplementary  Fig 10). 3.2 × 10 7 PFU of TBEV-FE (with a volume of 50 μl of virus suspensions in PBS) was administered via an intraperitoneal route to four-week-old WT or glp2r -/male mice. Mice were then monitored daily for body weight, survival and signs of pathogenesis. Mortality rate was assessed at the indicated time points. Animals were euthanized at 3 d, 7 d and 14 d post-infection, whole blood was obtained by cardiac puncture. Simultaneously, the brain, small intestine and spleen samples were collected after perfusion with PBS. The TBEV RNA and titers in serum and tissue samples were determined. The brain, small intestine and spleen samples were xed with 4% paraformaldehyde, embedded in para n and cut into 3.5-mm sections. Fixed tissue samples were used for hematoxylineosin (H&E) staining. The image information was collected using a Pannoramic MIDI system (3DHISTECH, Budapest).
Statistical analyses. The data were presented as the mean ± s.d. using Prism Version 6 (GraphPad). Statistical signi cance between groups were evaluated using Student's t-test or an analysis of variance (ANOVA) as indicated with 95% con dence intervals.
Reporting Summary. Further information on research design is available in the Nature Research Reporting Summary linked to this paper.
Data Availability. The Supplementary Excels provide data for the CRISPR-Cas9 screen and NGS gene analysis. The remainder of the data that support the ndings of this study are available from the corresponding authors upon reasonable request.