Attenuation of Getah Virus by a Single Amino Acid Substitution at Residue 253 of the E2 Protein that Might Be Part of a New Heparan Sulfate Binding Site on Alphaviruses

ABSTRACT

The emergence of new epidemic variants of alphaviruses poses a public health risk. It is associated with adaptive mutations that often cause increased pathogenicity. Getah virus (GETV), a neglected and re-emerging mosquito-borne alphavirus, poses threat to many domestic animals and probably even humans. At present, the underlying mechanisms of GETV pathogenesis are not well defined. We identified a residue in the E2 glycoprotein that is critical for viral adsorption to cultured cells and pathogenesis in vivo. Viruses containing an arginine instead of a lysine at residue 253 displayed enhanced infectivity in mammalian cells and diminished virulence in a mouse model of GETV disease. Experiments in cell culture show that heparan sulfate (HS) is a new attachment factor for GETV, and the exchange Lys253Arg improves virus attachment by enhancing binding to HS. The mutation also results in more effective binding to glycosaminoglycan (GAG), linked to low virulence due to rapid virus clearance from the circulation. Localization of residue 253 in the three-dimensional structure of the spike revealed several other basic residues in E2 and E1 in close vicinity that might constitute an HS-binding site different from sites previously identified in other alphaviruses. Overall, our study reveals that HS acts as the attachment factor of GETV and provides convincing evidence for an HS-binding determinant at residue 253 in the E2 glycoprotein of GETV, which contributes to infectivity and virulence.

IMPORTANCE Due to decades of inadequate monitoring and lack of vaccines and specific treatment, a large number of people have been infected with alphaviruses. GETV is a re-emerging alphavirus that has the potential to infect humans. This specificity of the GETV disease, particularly its propensity for chronic musculoskeletal manifestations, underscores the need to identify the genetic determinants that govern GETV virulence in the host. Using a mouse model, we show that a single amino acid substitution at residue 253 in the E2 glycoprotein causes attenuation of the virus. Residue 253 might be part of a binding site for HS, a ubiquitous attachment factor on the cell surface. The substitution of Lys by Arg improves cell attachment of the virus in vitro and virus clearance from the blood in vivo by enhancing binding to HS. In summary, we have identified HS as a new attachment factor for GETV and the corresponding binding site in the E2 protein for the first time. Our research potentially improved understanding of the pathogenic mechanism of GETV and provided a potential target for the development of new attenuated vaccines and antiviral drugs.

INTRODUCTION

Getah virus (GETV) is an enveloped, single-stranded, positive-sense RNA virus belonging to the Alphavirus genus of the family Togaviridae. The genomic RNA of the virus has two open reading frames that encode the four nonstructural proteins (nsP1-nsP2-nsP3-nsP4) and the structural proteins (capsid [C]-E3-E2-6K-E1). The glycoproteins E1 and E2 form a spike in the viral envelope, which plays an important role in virus binding to cell surface receptors and membrane fusion. GETV, a re-emerging mosquito-borne virus, has a broad geographical distribution, including Malaysia, Japan, Korea, China, Southeast Asia, Russia, Australia, and so on (1). GETV has been shown to infect various hosts, such as horses, pigs, cattle, blue foxes, kangaroos, humans, monkeys, and birds (2). Infected horses exhibited pyrexia, rash, and edema in the hind legs, while the clinical symptoms of pigs with GETV infection often are abortion and sow reproductive disorders (2, 3). Horses and pigs may play an important role in amplification in the natural transmission cycle of GETV (4). So far, GETV has been detected in various animals distributed in 15 provinces of China (5). Alphaviruses transmitted by arthropods are rapidly emerging and re-emerging human pathogens and continue to pose a global threat. Alphavirus can cause serious diseases in animals and humans, especially current epidemic strains revealed increased virulence caused by genetic mutations. Chikungunya (CHIKV), Ross River (RRV), O’nyong nyong virus (ONNV), and Mayaro viruses (MAYV) cause acute and chronic arthritis affecting millions of people globally, and Venezuelan equine encephalitis virus (VEEV), Eastern equine encephalitis virus (EEEV), and Western equine encephalitis virus (WEEV) cause fatal encephalitis, which is associated with significant morbidity and/or mortality. The lack of vaccines or antiviral measures approved for use in humans is a major obstacle to the treatment of future epidemics. It is important to carefully monitor alphaviruses evolution to prevent further public health issues.

Several alphaviruses utilize heparan sulfate (HS), a negatively charged glycosaminoglycan (GAG) widely present on the cell surface or extracellular matrix proteins (6). As an attachment factor, HS is recognized by either the wild-type virus or by variants that were serially passaged on cell lines expressing HS abundantly (7). A strange feature of HS-mediated attachment is that it can increase the infectivity of alphavirus in vitro, but it can enhance or decrease virulence in vivo, depending on the virus and inoculation route (8–10).

Amino acids that are involved in HS-binding have been identified for several alphaviruses. Natural isolates of North American eastern equine encephalitis virus (NA-EEEV) with high neurovirulence use HS as an attachment factor (11). Residues 71, 74, and 77 on E2 (E2-71, E2-74, and E2-77) are the main binding sites for HS, and mutations of these residues reduce the infectivity of cells in vitro but do not diminish the ability of virus replication in vivo (11). E2-71 and E2-74 are critical amino acids for HS binding, and they display cooperativity. EEEV strains with a stronger affinity for HS are less pathogenic in mice after subcutaneous (s.c.) inoculation but exhibit higher neurotoxicity after intracranial (i.c.) injection (8).

The exchange of Gly82Arg, a residue that lies on the molecule's surface in E2 of CHIKV, results in enhanced HS affinity (12). The mutation was acquired upon serial passage of the virus in cell culture. However, this exchange also contributes to the attenuation of CHIKV by neutralizing antibodies targeting E2 domain A and immune-mediated clearance (13). Small-plaque mutants of the Sindbis virus (SINV) with positively charged amino acid substitutions in E2 exhibit increased cell infectivity (14). It is speculated that stronger HS binding enhances the clearance rate in s.c. inoculated mice and weakens the viral diseases that rely on high-titer viremia. However, i.c. injection can increase neurovirulence and morbidity and/or mortality (9). In RRV, amino acid substitutions in E2-218 residue reportedly determined HS interactions and enhanced replication ability on avian cells and BHK-21 cells (15, 16).

VEEV has emerged repeatedly, infecting tens of thousands of people (17–19). Sequence analyses showed that amino acid substitutions resulting in increased positive charges on the surface of E2 were associated with several epizootic emergence events (20, 21). Substitutions in the E2 glycoprotein from glutamic acid or threonine to lysine increased the positive charge on the molecule's surface and increased adsorption and invasion of cells in vitro. Adaptive mutations with stronger affinity to HS showed attenuated virulence in s.c. inoculated mice, similar to SINV (10). The appearance of VEEV serotypes IAB and IC was associated with the change of residue 213 from threonine to either lysine or arginine, and these viruses exhibited higher viremia titers and replicated efficiently in equines. This site is also affected by positive selection and convergent evolution of the epizootic phenotype (22). Thus, investigating the effects of positively charged E2 amino acids on epizootic phenotypes and continuously monitoring viral alterations in natural circulation are very important to prevent epidemic breakouts.

The increasing emergence of GETV poses a serious threat to animal health and public health. However, the etiology and pathology of GETV are largely unclear. In the past few decades, a large number of careful studies have explored the molecular mechanisms of infectivity for a variety of alphaviruses, including adsorption, internalization, and cell entry (23). In the current study, we found that GETV utilizes HS for efficient attachment to target cells. Arginine at residue 253 increased affinity with HS and allows more efficient viral replication in vitro. In addition, we also found that E2-K253R attenuated the virulence of mice, infected either i.c. or s.c., due to enhanced clearance from the blood. These findings might contribute to the development of attenuated vaccines.

RESULTS

E2 residue 253 plays a major role in determining the virulence of GETV.

According to our lab's previous epidemiological investigations, two amino acids, lysine and arginine, exist at position 253 of the GETV E2 protein in various GETV isolates. Computational analysis indicated that the change from lysine to arginine at E2 position 253 was caused by positive selection of the epizootic phenotype (He et al., unpublished). Basic amino acids in close proximity might promote binding to heparan sulfate (HS), which enhances replication of several alphaviruses in cell culture (24–26). Analogously, the acquisition of an arginine residue is partially responsible for the attenuation of some alphaviruses (12). To investigate the effect of the amino acid substitution Lys253Arg in vivo and in vitro, we rescued the wild-type virus, named rGETV-HN, and the mutant virus, E2-K253R and E2-K253A, with a reverse genetics system (Fig. 1). Although an adult mouse model is commonly used in alphaviruses studies (27), considering the lethal effect of GETV on piglets (28), we chose 2-day-old mice to conduct virulence experiments because GETV can cause the death of suckling mice (29).

FIG 1.

Assembly of the GETV genome into a pSMART-LCKAN plasmid. Scheme of the plasmid used to rescue recombinant Getah virus. The pSMART-LCKAN vector and the sequence of the GETV-HN strain were screened for unique restriction sites, and the seven sites shown in the schematic were selected for cloning. Fusion PCR was used to introduce the cytomegalovirus (CMV) promoter sequence at the 5' end and HDVRz and BGH sequences at the 3′ end of GETV-HN strain. By making synonymous mutation in the E3 gene (primers are shown in Table 1), a XhoI site in the genome was silenced as the genetic marker to rescue the virus, and another XhoI restriction site was introduced at the CMV 5′ end. The full-length infectious cDNA clone was assembled by sequentially cloning the fragments A to F amplified by PCR into the pSMART-LCKAN plasmid using the indicated restriction sites. Upon transfection into cells, viral RNA is transcribed under the control of the CMV promoter, which serves as both genomic and viral mRNA. Two open reading frames encode the viral nonstructural proteins nsP1, nsP2, nsP3, and nsP4, and the viral structural proteins capsid (C), E3, E2, 6K, and E1. UTR, untranslated regions. HDVRz, hepatitis delta virus ribozyme. BGH, transcription terminator from the bovine growth hormone. Length is proportional.

To investigate whether residue 253 in E2 is a virulence determinant, we infected 2-day-old suckling mice, either s.c. or i.c. with 10⁶/TCID₅₀ of rGETV-HN or E2-K253R. When injected subcutaneously, all the mice infected with rGETV-HN succumbed to the infection within 2.5 days postinfection, and the average survival times (ASTs) were 2 days (Fig. 2A). All of them died before the onset of severe weight loss in the s.c. group (Fig. 2C). No deaths occurred in the groups infected with the E2-K253R, similar to the uninfected control group (Dulbecco's modified Eagle's medium [DMEM]) (Fig. 2A). In the E2-K253R and control group, none of the infected mice exhibited neurological or other symptoms, such as weight loss, regardless of the inoculation routes. However, the weight of E2-K253R increased slowly compared with the control group (Fig. 2C and D). In contrast, there are 85.7% mortality in suckling mice infected with rGETV-HN by the i.c. route (Fig. 2B). Only one mouse survived and gained weight slowly in the i.c. group (Fig. 2D). The ASTs of rGETV-HN were 2 days in i.c. group (Fig. 2B). These results demonstrate that the K253R mutation in the E2 protein confers dramatically reduced viral virulence and prevents lethal infection with GETV.

FIG 2.

Effects of mutation at residue 253 for GETV virulence in mice infected by the intracranial and subcutaneous route. Groups of seven 2-day-old ICR mice were infected with 25 μL of each virus containing 10⁶/TCID₅₀, either s.c. (A, C) or i.c. (B, D). Groups injected with medium (DMEM) were used as control. Mice were observed until day 14. Weight changes (C to D) and death of mice (A to B) were recorded at 12-h intervals. A significant difference in survival (***, P < 0.001) only occurred between-group rGETV-HN or E2-K253R and DMEM.

E2 residue 253 impairs GETV replication in vivo.

To assess the effect of residue 253 on viral replication in vivo, we determined viral loads by TCID₅₀ in different tissues (knee joint, brain, kidney, trachea, turbinate, and lung) of suckling mice at 1 and 2 days after inoculation by either s.c. or i.c. route. The virus titer of E2-K253R was significantly lower than that of rGETV-HN in all organs and at all time points (Fig. 3). The values were statistically significantly different, except for brains, trachea, and lung samples 24 h after s.c. inoculation, suggesting that the E2-K253R mutation reduces GETV's ability to replicate in vivo.

FIG 3.

Effect of mutation at residue 253 on replication of GETV in multiple tissues of mice. Two-day-old mice (three per group) were infected with 25 μL of each virus containing 10⁶/TCID₅₀, either s.c. or i.c. Central nervous system tissue (brain) and extraneural tissues (knee joint, kidney, trachea, turbinate, and lung) were collected at 24-h and 48-h postinfection (p.i.) and virus titers were determined by TCID₅₀. Results are presented as mean ± SD, and the asterisks indicate significant differences compared to rGETV-HN-infected mice. (Student's t test, *, P < 0.1, **, P < 0.01; ***, P < 0.001).

E2 residue 253 is a determinant of GETV infectivity in mammalian cell culture.

Next, we compared the attachment of rGETV-HN and E2-K253R to mammalian cells. BHK-21, PK-15, N₂A, or U251 cells were incubated with GETV at a multiplicity of infection (MOI) of 1 for 1 h at 4°C, and virions adsorbed to the cell surface were directly quantified by qRT-PCR. We found that compared with rGETV-HN, 2 to 10 times more particles of the variant E2-K253R were attached to all cell lines (Fig. 4A to D).

FIG 4.

E2 residue 253 is a determinant of GETV infectivity in mammalian cell culture. (A to D) Pre-chilled BHK-21 (A), PK-15 (B), N₂A (C), and U251 (D) cells were incubated with the precooled rGETV-HN or E2-K253R (MOI = 1). After adsorbing for 1 h at 4°C, cells were washed with ice-cold PBS 3 times on ice. Subsequently, cells were lysed with TRIzol for qRT-PCR. (E to F) E2 residue 253 mediates direct interaction with HS. Approximately 10⁶/TCID₅₀ of rGETV-HN and E2-K253R virions were incubated with 30 μL of washed heparin- or His-conjugated magnetic beads at 4°C for 1 h. Beads were washed with the binding buffer 3 times, and bound virus particles were eluted with a high salt buffer. (E) An equal amount of eluate, as well as input virus, were subjected to Western blot analysis. (F) Quantification of band intensities by image J. The percentage of virus bound to beads was quantified by optical densitometry and calculated as the ratio of eluted virus particles to input virus particles. The quantification is the result of three independent experiments. (G) Pre-chilled BHK-21 and PK-15 cells were incubated with the precooled rGETV-HN or E2-K253A (MOI = 1) for 1 h at 4°C, and attachment experiments were carried out as described in A to D. Data were expressed as mean ± SD derived from three infected cell cultures. (Student's t test, *, P < 0.01; ***, P < 0.001).

We next explored whether rGETV-HN binds to heparin and whether binding increased in E2-K253R. Identical genome copy numbers of rGETV-HN or E2-K253R were incubated with heparin-conjugated magnetic beads. Bound GETV particles eluted by high salt buffer were subjected to immunoblotting using pAbs against the viral capsid protein (cap). The blot shows that bands representing E2-K253R particles eluted from the heparin-conjugated magnetic beads were roughly 2 times stronger than bands corresponding to rGETV-HN particles with a Lys at position 253. (Fig. 4E). Quantification of band densities revealed that 80% of input E2-K253R particles were bound to heparin-conjugated magnetic beads, but only 35% of rGETV-HN particles. None of the viruses bound to His-conjugated magnetic beads which were used as a negative control (Fig. 4F).

In addition, we mutated residue 253 from Lys to an uncharged Ala to investigate whether a positive charge positively affects virus binding to the cell surface. BHK-21 and PK-15 cells were incubated with rGETV-HN (MOI = 1) or with E2-K253A for 1 h at 4°C, and virions adsorbed to the cell surface were directly quantified by qRT-PCR. The infectivity of E2-K253A was significantly reduced to 20% to 30% compared with rGETV-HN (Fig. 4G). These results indicate that the presence of basic amino acid at residue 253 of E2 plays an important role in binding of GETV to cells.

Glycosaminoglycans neutralize infectivity of several GETV strains in a dose-dependent manner.

To assess whether soluble GAGs act as a competitive inhibitor and block the infectivity of GETV, we performed competition assays using heparin (Hp) or chondroitin sulfate (CS). First, we determined their non-hazardous concentration and found that Hp or CS at 20 mg/mL did not affect the viability of BHK-21, PK-15, and U251 cells (Fig. 5A to C). Preincubation of GETV-GX with Hp or CS (1 mg/mL) for 30 min at 37°C was done to neutralize viral infectivity, measured as viral mRNA levels by qRT-PCR (Fig. 5D to F) and infectious particles released from cells by TCID₅₀ (Fig. 5G to I). The effect of Hp was stronger in all cell lines, but incubation with CS also exhibited a significant effect in PK-15 cells.

FIG 5.

Glycosaminoglycans neutralize infectivity of several GETV strains in a dose-dependent manner. (A to C) BHK-21, PK-15, and U251 cells were incubated with 0 to 20 mg/mL of Hp or CS for 1 h. Cell viability was evaluated by the CCK-8 Kit according to the manufacturer’s instructions. (D to I) GETV-GX (MOI = 0.1) was preincubated at 37°C for 30 min with 1 mg/mL Hp. The virus incubated with BSA was the control. Cells were infected with this mixture for 1 h at 37°C and then washed with PBS 3 times to remove unbound virions. Maintenance medium was then added, and cells continued to cultivate until 24-h postinfection. (D to F) Cells were lysed, and E2 mRNA levels were determined by qRT-PCR. (G to I) Viral titers in supernatants after Hp or CS treatment were measured using TCID₅₀. (J to N) Three strains of GETV-GX, GETV-HN, and GETV-FJ were incubated with increasing concentrations of Hp (0, 0.125, 0.25, and 0.5 mg/mL) for 30 min at 37°C prior to infection of BHK-21 cells. After infected 1 h at 37°C, cells were washed 3 times with PBS. (J) Cells were stained with GETV pAbs of the capsid (cap). Green represents virions, and blue is DAPI. Scale bar, 500 μm. (K) The green fluorescence intensity in J was quantified with the NIS-Elements AR Analysis 4.51.00 (Nikon). Each captured frame was processed to remove the background. Quantification of TGEV-infected cells from the immunofluorescence assay (IFA) images was normalized to untreated control cells. (L to N) Viral titers in supernatants were measured by TCID₅₀. All data were pooled from three independent biological replicates and presented as mean ± SD. Bars represent standard deviations, and the asterisks indicate significance relative to controls (Student's t test, *, P < 0.1; **, P < 0.01; ***, P < 0.001).

We also tested whether other GETV strains were sensitive to Hp. GETV-GX, GETV-HN, and GETV-FJ, all possess a lysine at position 253, were preincubated with increasing concentrations of Hp (0.125, 0.25, and 0.5 mg/mL) for 30 min at 37°C before infecting the cells. Successful infections of cells were determined by immunofluorescence using pAbs against the E2 protein (Fig. 5J). Measuring the green fluorescence intensity revealed a decrease with increasing Hp concentration (Fig. 5K). The number of infectious virus particles released into the supernatant also decreased in a dose-dependent manner (Fig. 5L to N). We found that all strains are sensitive to Hp in all three assays, although the inhibitory efficiency of Hp varies among different strains.

Inhibition of GETV by glycosaminoglycans occurs at the attachment stage.

Similar assays were performed to determine whether Hp and CS mainly affect the adsorption process between the three GETV strains and cells. GETV-GX, GETV-HN, and GETV-FJ were preincubated with either Hp (1 mg/mL) or CS (1 mg/mL) prior to infection of BHK-21 or PK-15 cells for 1 h at 4°C, and virions adsorbed to the cell surface were directly quantified by qRT-PCR. Consistent with our above results (Fig. 5D to I), incubation of GETV-HN with Hp prior to attachment resulted in almost complete inhibition of GETV attachment by comparing relative viral genomes, whereas preincubation with CS showed a poor inhibitory effect in BHK-21 or PK-15 cells (Fig. 6B and E). The semblable trend was observed for the other two strains GETV-GX and GETV-FJ (Fig. 6A, D, C, and F).

FIG 6.

Inhibition of GETV by heparin occurs at the attachment stage. (A to I) Three strains of GETV (MOI = 1 in attachment assay, MOI = 0.01 in plaque assay) were incubated for 30 min at 37°C with Hp or CS (1 mg/mL in attachment assay, 0.1 mg/mL in plaque assay). NC, BSA-treated cells. Mock, uninfected cells. Pre-cooled (10 min) cells were then incubated with the mixtures for 1 h at 4°C. Unabsorbed virions were removed by washing with ice-cold PBS. (A to F) Cells were lysed and analyzed. Determination of relative viral genome numbers by qRT-PCR. (G to I) Cells were covered with methylcellulose and incubated for 48 h before staining to visualize plaques. The number of plaques was counted normalized to control cells (NC). Results are the mean ± SD of three experiments, and the asterisks indicate significant differences by Student's t test compared with BSA-treated cells (NS, no significant difference; *, P < 0.1; ***, P < 0.001).

We also further verified this result by determining the inhibitory effect of Hp on plaque formation in BHK-21 cells using previously reported protocols (30). GETV was incubated with Hp or CS at 0.1 mg/mL concentration for 30 min at 37°C before infecting BHK-21 cells. Cells were then overlaid with methylcellulose and incubated further for 24 h. The number of plaques as a surrogate marker for attachment of infectious virus particles was reduced up to 90%, whereas CS at the same concentration showed a reduction of only∼70% in GETV-HN strain or no effect in GETV-GX and GETV-FJ strains (Fig. 6G to I). Therefore, inhibition of GETV by glycosaminoglycans occurs at the attachment stage.

GETV infection strongly depends on the sulfation of heparan sulfate.

NaClO₃ inhibits adding O-sulfate groups to GAGs, while Na₂SO₄ can restore this process (31, 32). To investigate whether GETV requires highly sulfated HS polysaccharides for attachment, BHK-21 or PK-15 cells were passaged in the presence of 0, 25, or 50 mM NaClO₃ for 48 h prior to infection. Importantly, we did not observe any apparent changes in cell morphology of BHK-21 or PK-15 cells cultured for 3 days in the presence of up to 50 mM NaClO₃ (data not shown). qRT-PCR performed after 24-h postinfection with GETV-GX (MOI = 0.1) revealed a decrease in GETV E2 mRNA by 80% to 90% in cells treated with NaClO₃ compared with the untreated cells (Fig. 7A and B). The detection of E2 and cap proteins by Western blot (Fig. 7C and D) and the determination of virus titers in the supernatant by TCID₅₀ (Fig. 7E and F) also showed the same decreasing trend. Supplementation of NaClO₃-treated cells with 0.8 mM sodium sulfate (Na₂SO₄) or just incubation in medium 48 h before GETV infection greatly rescued the inhibitory effect of NaClO₃ in all experiments. Furthermore, treatment of BHK-21 cells with NaClO₃ significantly reduced the attachment (visualization through plaques) of three GETV strains (MOI = 0.01) normalized to untreated controls in a dose-dependent manner (Fig. 7G to I).

FIG 7.

GETV infection strongly depends on the sulfation of heparan. (A to I) BHK-21 and PK-15 cells were treated with 25 mM or 50 mM NaClO₃ for 48 h. Cells were then divided into three parts, one part was treated with 25 mM or 50 mM NaClO₃, one part with 0.8 mM Na₂SO₄, and the other was cultured without any treatment for 48 h. Cells were then infected with GETV-GX (MOI = 0.1) and analyzed 24 h later. (A to B) E2 mRNA levels were evaluated by qRT-PCR. (C to D) The expression level of E2 and cap protein was analyzed by Western blot. (E to F) Virus titers in the supernatant were measured using TCID₅₀ in triplicate. (G to I) The same treatments were done before BHK-21 cells were incubated with the three strains of GETV (MOI = 0.01) for 1 h at 4°C. Cells were then covered with methylcellulose, incubated for 48 h, and stained to visualize plaques. The results are representatives of at least three independent experiments. The asterisks indicate significance relative to controls by Student's t test (*, P < 0.1; **, P < 0.01; ***, P < 0.001).

The ability of the virus to adsorb to the cells depends on the electrostatic attraction.

Inhibition of viral replication in BHK-21 or PK-15 cells by high concentrations of NaCl was investigated (8). First, we determined the non-hazardous concentration of NaCl and found that 300 mM NaCl had no negative effect on BHK-21 and PK-15 cell viability (Fig. 8A). We added increasing NaCl concentrations (0, 150, and 300 mM, final concentration) during the virus incubation with cells. GETV-GX virus infection was negatively correlated with ionic strength, measured as virus mRNA synthesized inside cells and infectious viral particles in supernatants (Fig. 8B to E). Furthermore, the addition of a high concentration of NaCl in BHK-21 cells significantly reduced cell attachment of GETV-GX, GETV-HN, and GETV-FJ, measured as plaque counts normalized to control (Fig. 8F to H). Given that a high concentration of NaCl buffer disrupts ionic interactions, we conclude that the attachment of GETV to host cells is mediated by electrostatic attraction.

FIG 8.

Virus adsorption to the cells depends on electrostatic attraction. (A) BHK-21 and PK-15 cells were incubated with increasing concentrations of NaCl, and cell viability was determined after 2 days. (B to E) Increasing concentrations of NaCl (0, 150, and 300 mM, final concentration) were present during the incubation period of cells with GETV-GX (MOI = 0.1) at 37°C for 1 h. After 24 h, the following analyses were done: (B to C) virus mRNA levels were determined by qRT-PCR, (D to E) virus titers present in the supernatant were determined by TCID_50. (F to H) Increasing concentrations of NaCl were mixed with three strains of GETV (MOI = 0.01), and cells were incubated with the virus for 1 h at 4°C. Cell monolayers were covered with methylcellulose after incubation, incubated for 48 h, and stained to visualize plaques. All results are from three independent experiments. Error bars indicate SD. ***, P < 0.001 (compared with 0 mM NaCl treated controls as determined by Student's t test).

E2-K253R is more sensitive to HS-binding than rGETV-HN.

Next, we compared the sensitivity of HS-binding of rGETV-HN and E2-K253R to NaCl. We added NaCl during the incubation period of virus with cells and determined relative virus genome copy numbers after attachment. We found that the number of E2-K253R virus particles adsorbed was reduced by∼90% compared with the untreated control group, whereas the rGETV-HN showed a reduction of only∼70%, 300 mM NaCl had a stronger inhibitory effect on E2-K253R compared with rGETV-HN, indicating that the E2-K253R is more dependent on the electrostatic attraction for efficient cell attachment (Fig. 9A). Likewise, E2-K253R attachment is also more sensitive to the preincubation of cells with different concentrations of NaClO₃. The number of adsorbed virions was reduced to 90%, 60%, and 45% relative to the untreated controls at a NaClO₃ concentration of 10 mM, 25 mM, and 50 mM, respectively, but the reduction was more pronounced measured as plaque counts, 75%, 35%, and 25%, respectively for E2-K253R. In short, compared with rGETV-HN, E2-K253R is more dependent on the sulfation of cell surface receptors to infect cells (Fig. 9B). These results suggest that GETV strains containing arginine at E2 residue 253 improved affinity with GAGs for efficient cell infection.

FIG 9.

E2-K253R is more sensitive to HS-binding than rGETV-HN. (A) NaCl (300 mM) was mixed with E2-K253R or rGETV-HN (MOI = 1) and adsorbed to BHK-21 cells at 4°C for 1 h. Cells were then washed and collected for qRT-PCR. (B) E2-K253R and rGETV-HN (MOI = 0.01) were adsorbed to BHK-21 cells pretreated with different concentrations of NaClO₃ (10, 25, and 50 mM) at 4°C for 1 h. Cells were washed and incubated at 37°C for 48 h and stained to visualize plaques. (C) Virus clearance in vivo. Virus titers in the serum were determined at 5, 10, 20, 40, and 60 min after intravenous tail injection of mice with 10⁶/TCID₅₀ of either E2-K253R or rGETV-HN. The titer at each time point was measured by TCID₅₀. All data were normalized to the untreated controls for three independent experiments performed in triplicate. Error bars indicate SD. ns means no significance; **, P < 0.01; ***, P < 0.001; ^##, P < 0.01; ^###, P < 0.001 (in comparison to untreated controls as determined by Student's t test).

To assess the contribution of E2 protein residue R253 to virus clearance from the blood, clearance kinetics were determined with rGETV-HN and E2-K253R in vivo according to a previously reported method (33). ICR male mice aged 5 weeks were infected by intravenous (i.v.) inoculation with a high virus dose (10⁶/TCID₅₀), and virus titers in the serum within 60 min after infection were determined by TCID₅₀ on BHK-21 cells. The E2-K253R, which showed an enhanced HS-binding phenotype, was cleared from the blood more rapidly (around 10⁴-fold reduction already 10 min after infection) than the rGETV-HN (around 10³-fold reduction after 10 min) (Fig. 9C). We speculate that the arginine at residue 253 of the GETV E2 protein has an enhanced ability to bind to GAGs, enhancing the clearance of the virus from the blood.

Localization of residue 253 in the E2/E1 structure suggests a new HS-binding motif in alphaviruses.

Because no structure of E1E2 is available for GETV, we used the Cryo-EM structure of VEEV, which is closely related to GETV, to highlight residue 253 (corresponding to residue 254 in VEEV) in the E1/E2 spike. It is located at the periphery of the molecule in a short β-sheet, which contains two other basic amino acids in close vicinity, Lys 252 and His 256 (Fig. 10A and B, red spheres). Their side chains point downwards toward the E1 subunit. Localized at short distance to residue 253 are three other basic amino acids in the E1 subunit, and their side chains point upwards toward the E2 subunit (Fig. 10C). Visualization of the electrostatic surface potential of this region shows that Lys 252 and Lys 254 in E2 and Lys 52 and Arg73 form a basic groove at the surface of the molecule (Fig. 10D). Basic residues at all these positions are present in GETV and other alphaviruses, such as CHIKV, RRV, EEEV, and MAYV. We thus suggest that these residues are a new HS-binding site, which is different from the two sites (an axial and three peripheral) already described for EEEV (Fig. 10A and B, magenta and cyan spheres). An exchange from Lys to Arg probably increases the basic strength of the molecule's surface, since arginine, with an isoelectric point of 10.76, is more basic than lysine, having an isoelectric point of 9.74.

FIG 10.

A putative HS-binding site in E1/E2 of GETV and other alphaviruses. (A to B) The side view and top view of the E1/E2 structure of VEEV show the HS-binding sites, which are highlighted as spheres. The red spheres indicate the putative HS-binding site in GETV and VEEV, the purple and cyan spheres indicate the known axial and peripheral HS-binding sites identified for EEEV. (C) Detail of the E1/E2 structure. E2 is in green and E1 in blue. Basic amino acids that might form an HS-binding site in GETV are shown as sticks and labeled. The dotted lines and the numbers indicate the distance in Å between residues. The alphaviruses GETV, CHIKV, RRV, EEEV, and MAYV, all contain basic amino acids at these residues. In the GETV-HN strain, residue 110 is not a Lys but an Arg. Note that residue 254 in VEEV corresponds to residue 253 in GETV. (D) The electrostatic surface potential of (C). Basic regions are colored blue, and acidic regions are red. (E to G) Ionic interactions between two β-strands in E2, one of which contains the basic residues of the proposed HS-binding site. (E) In VEEV, Lys254 forms an ionic bond with Glu 166 located at the parallel β-sheet. (F) In the GETV-HN strain, Glu 166 is replaced by Asp, which increases the length of the ionic bond. (G) Exchange of Lys254 by Arg in E2-K253R would decrease the length of the ionic bond. The dotted lines and the numbers indicate the distance in Å between residues, which is 3.4 Å in (E), 4.0 Å in (F), and 2.9 Å in (G). All figures were created with PyMol from pdb-file 3J0C.

In addition, in VEEV, the side chain of Lys 254 forms an electrostatic interaction with Glu166, which is located in the antiparallel β-strand (Fig. 10E). Residue 166 is always an acidic amino acid in the alphaviruses mentioned above, but in GETV, Glu is replaced by Asp. Because Asp has a shorter side chain than Glu, the distance of the ionic interaction increases from 3.4 to 4.0 Angström, and hence its strength decreases (Fig. 10F). We also exchanged Lys by Arg using the mutagenesis function of PyMol. Although 23 different conformational isomers of the side chain of Arg (rotamers) exist, only one does not lead to an obvious clash with side chains of other amino acids. This Arg rotamer would again decrease the distance to Asp166 to 2.9 Å (Fig. 10G), and hence the exchange of Lys by Arg in E2-K253R might subtly affect the geometry of the binding groove and its affinity for HS.

DISCUSSION

Cell attachment is the initial and essential step of virus infection. So far, no specific molecules and proteins have been identified as attachment factors for GETV infection. Here, we discovered for the first time that GETV uses HS as an attachment factor. What is noteworthy is that HS may simply concentrate virus particles at the cell surface before interacting with other more specific unknown receptors. Our conclusion is based on the following lines of evidence: GETV binds specifically to beads coupled to Hp (Fig. 4E and F). In cell culture, virus attachment is markedly inhibited by soluble Hp (Fig. 5), and pretreatment of cells to inhibit the synthesis of highly sulfated HS effectively reduced GETV infection (Fig. 7). In addition, GETV adsorption to cells is blocked by a high salt buffer (Fig. 8), which prevents the electrostatic interaction between virus and HS.

Previous studies have shown that in NA-EEEV, amino acid mutations in natural circulating strains allow the virus to use HS as an attachment factor (8). Likewise, the interaction of E2 glycoprotein with HS is crucial for cellular infection of the cell culture-adapted SINV strain through selective mutation for positively charged amino acids (34, 35). We also provide evidence that residue 253 in the glycoprotein E2 is involved in binding to HS. In previous studies, we compared the sequence of emerging GETV and found a Lys to Arg mutation at the residue 253 of the GETV E2 protein (He et al., unpublished). Positive selection pressure may have acted on residue 253 of E2 causing the change from Lys to Arg (He et al., unpublished), and it was speculated that mutations might change the ability to interact with HS. To explore the impact of these two amino acids, we established a reverse genetics system for GETV for the first time and rescued the wild-type rGETV-HN and the mutant virus E2-K253R and E2-K253A. Our experiments confirmed that when the E2 residue 253 is arginine, the virus has a higher infectious ability in mammalian cells and binds stronger to beads coupled to Hp (Fig. 4). In addition, exchange of Lys by Ala reduced attachment of virus to the cell surface. Furthermore, E2-K253R showed stronger sensitivity to electrostatic interference and sulfation degree of HS than rGETV-HN (Fig. 8 and 9).

Most importantly, the mutant E2-K253R shows lower pathogenicity in a mouse model and replicates to lower titers in all tissues of experimentally infected mice (Fig. 2 and 3). Furthermore, the E2-K253R has a faster clearance rate from the blood than the rGETV-HN (Fig. 9C). We speculate that the weaker virulence of E2-K253R is due to the faster clearance rate, which might be due to stronger binding to GAGs. These findings suggest that the mechanism of virulence attenuation of GETV in vivo closely resembles that proposed previously for GAG-binding variants of other alphaviruses or flaviviruses (33).

The general HS recognition amino acid sequence is XBBXBX or XBBBXXBX, where B means basic and X any amino acid (36). Residue 254 in E2 of closely related VEEV (corresponds to 253 in GETV) is located at the periphery of the E1/E2 spike in close proximity to two other basic amino acids, Lys 252 and His 256, and thus conforms to the consensus sequence. The side chains of these amino acids point downwards toward the E1 subunit, which contains three other basic amino acids, and their side chains point upwards toward the E2 subunit (Fig. 10). Basic residues are conserved at these positions in GETV and other alphaviruses, such as CHIKV, RRV, EEEV, MAYV. Thus, these amino acids might represent a new HS-binding site, different from sites previously identified in other alphaviruses. Molecular docking experiments predicted that CHIKV binds to HS by a positively charged pocket on the E2-E1 spike, which is also present in other alphaviruses, such as SINV and VEEV (37). The binding sequence motif XBXXBX is part of the EF loop of domain A, which is highly conserved in sequence and structure in alphaviruses (38). The 71 and 74 lysines of EEEV create a positively charged site toward the upper edge of the gap between domains A and B through which the E1 fusion loop passes (8). Positively charged residues near this region may also play an important role in HS binding for SINV (E2 70K) (39), VEEV (E2 76K) (10), and CHIKV (E2 82R) (12).

The attenuation of viruses due to enhanced binding to HS has been reported before for other alphaviruses but occurred upon replacing an uncharged by a basic, positively charged amino acid, usually by an arginine. The substitution E2-Lys253Arg does not introduce a positive charge, but the basic strength of arginine is stronger than the strength of lysine. Besides, the tighter interaction observed for E2-K253R may result from stronger hydrogen bond formation between the guanidino group of arginine and the sulfate of HS (40, 41). Furthermore, the side chain of Lys254 in VEEV and probably Lys253 in GETV form an ionic bond with a negatively charged residue located in an antiparallel β-sheet. The exchange Lys to Arg might strengthen this interaction and optimize the geometry of the binding pocket for better HS-binding (Fig. 10E to G).

GETV is a neglected virus that has the potential to cause epidemics in animals and might also infect humans, thus causing a public health risk. Therefore, the molecular evolution of GETV should be continuously monitored. Here, we report that the substitution of arginine at E2 can significantly reduce virulence and, at the same time, increase binding to HS, an attachment factor for GETV. We speculate that the weaker virulence of E2-K253R is related to the faster virus clearance rate from the blood, perhaps by increased binding to GAG, as proposed previously for GAG-binding variants of other alphaviruses (33, 42, 43). Our data also provides a basis for the development of attenuated vaccines. Previous studies have shown that HS mimetics, such as pentosan polysulfate (PPS) and PG545, can be used to prevent and treat alphavirus infections, such as RRV and CHIKV (44). Our data support the potential of PPS and PG545 to also be effective against GETV infections.

MATERIALS AND METHODS

Cell and virus.

Baby hamster kidney (BHK-21) and African green monkey kidney (Vero) cell lines were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco by life) supplemented with 10% (vol/vol) fetal bovine serum (FBS, Biological Industries). Porcine kidney (PK-15), and Neuro-2A murine neuroblastoma cells (N₂A) were maintained in DMEM with 10% FBS (Sigma). The human glioma U251 cell line was passaged with minimum Eagle’s medium (MEM, Gibco by life) and 10% FBS (Sigma). All cells were incubated at 37°C in a humidified incubator with 5% CO₂. A maintenance medium with 2% FBS was used for infected cells.

Three GETV strains, named GETV-GX (MZ736796), GETV-HN (MZ736801), and GETV-FJ (MZ736799), were isolated from pigs and stored in our laboratory. Sequencing showed that the three strains exhibit some differences in amino acids, but each isolate possesses a conserved lysine at residue 253. Viruses were grown and titrated on Vero cells.

Establishing a reverse genetic system for GETV.

The reverse genetic system of rGETV-HN was successfully constructed for the first time. Based on the full-length sequence of GETV-HN, the genome was divided into six fragments amplified by PCR, and using a cDNA copy of the rGETV-HN cDNA as a template. Fragment A1 and fragment A2 (see Table 1) were amplified by conventional PCR, and fragment A was obtained by fusion PCR. Specific primers were used to amplify fragments B, C, D, and E. The same reverse primers of F were used to amplify four small fragments, F1, F2, F3, and F4, and finally, the fragment F was obtained by fusion PCR. Finally, a tail was added by fusion PCR to ensure that the virus could be reversed. The primers (Table 1) contained unique restriction sites that were used to ligate the six gene fragments, which were cloned into the low-copy-number plasmid pSMART-LCKAN step by step (Fig. 1). The E2-K253R and E2-K253A mutations were constructed by site-directed mutagenesis (primers are shown in Table 1), as previously described (45). DNA fragments which 13-bp overlap sequences for recombination were amplified using high amplification fidelity enzyme (Vazyme) with the primers specified in Table 1 and rGETV-HN as the template. According to the manufacturer's instructions, the resulting target DNA fragment and vector were digested with BstBI and MluI restriction enzymes, recovered from agarose gel with Cycle-Pure Kit (OMEGA) and ligated with T4 DNA Ligase (Vazyme) at 16°C for 10 h. According to the manufacturer’s instructions, ligation products were transformed into Top 10 competent cells, plated on solid Luria-Bertani (LB) medium containing kanamycin. After 12 h at 37°C, a single colony was picked, amplified in liquid LB medium at 37°C for 12 to 14 h, and plasmids were sequenced. Plasmids with the correct sequence were extracted from bacteria using the Endotoxin-free plasmid minikit I (OMEGA). The concentration of the plasmid was determined.

TABLE 1.

Sequences of the primers

Gene	Sequence (5′–3′)
GETV-A1	Forward: CCCAAGCTTGGGCCGCTCGAGGACATTGATTATTGACTAGT TATTAATAGTAATCA
	Reverse:GTCACACGTCCGCCATAGAGCTCTGCTTATATAGACCTCCCA
GETV-A2	Forward:TGGGAGGTCTATATAAGCAGAGCTCTATGGCGGACGTGTGAC
	Reverse:AACACACCAATGATCGTCGTGTGGT
GETV-B	Forward:GTGTTAACAGGGGACCTAATCAATCC
	Reverse:TTAGTGCGTATGCTGTCCATCGTC
GETV-C	Forward:TTATGGCCAAGACTCCAAGATGCT
	Reverse:CGTCGTTATGTTTTCGGTGGTGAT
GETV-D	Forward:ACTCTGCGGTGTTTAACGTAGAAAGC
	Reverse:GCGGACCATTCTTCTGTTCCTTCT
GETV-E	Forward:ACTACAATTGGCATCACGGTGCAGT
	Reverse:CTGTCTTGATGTCCACAGCTGCCT
GETV-F1	Forward:GGGTCGGCATGGCATCTCCACCTCCTCGCGGTCCGACCTGGGCATCCGAAG
	Reverse:TGCTCTAGACCATAGAGCCCACCGCATC
GETV-F2	Forward:CCTGGGCATCCGAAGGAGGACGTCGTCCACTCGGATGGCTAAGGGAGAG
	Reverse:TGCTCTAGACCATAGAGCCCACCGCATC
GETV-F3	Forward:TGGCTAAGGGAGAGCTCGGATCCGCCTCGACTGTGCCTTCTAGTTGCC
	Reverse:TGCTCTAGACCATAGAGCCCACCGCATC
GETV-F4	Forward:ACGACCTGCTCGAAGCCACGAT
	Reverse: GAGATGCCATGCCGACCCTTTTTTTTTTTTGTAAAATATTAAAAAAACAAATTAGACGCC
GETV-E-XhoI-Mut	Forward:ACGACCTGCTCGAAGCCACGAT
	Reverse:TCGAGCAGGTCGTAGTAGCCCGG
E2-K253R	Forward:GTCTCGCAAAGGTAGAGTGCACGTACCTTTCCCTCTGACC
	Reverse:ACCTTTGCGAGACAACTGGTCGGCTC
E2-K253A	Forward:GTCTCGCAAAGGTGCAGTGCACGTACCTTTCCCTCTGACC
	Reverse:ACCTTTGCGAGACAACTGGTCGGCTC
β-actin-mouse	Forward:GGTGGGAATGGGTCAGAAG
	Reverse:AGCTCATTGTAGAAGGTGTGG
β-actin-pig	Forward:CTCCATCATGAAGTGCGACGT
	Reverse:GTGATCTCCTTCTGCATCCTGTC
GAPDH-human	Forward:GCACCGTCAAGGCTGAGAAC
	Reverse:TGGTGAAGACGCCAGTGGA
GETV-E2	Forward:AGAGAACTGACGGTGAAACTG
	Reverse:CCATCTGTACTCGATCCCTTC

For recombinant virus rescue, plasmids (2 μg) encoding rGETV-HN, E2-K253R, E2-K253A, and empty plasmid as an internal control were transfected into BHK-21 cells using Lipofectamine 2000 (Thermo) according to the manufacturer's instructions. Cells were cultured at 37°C and monitored until 80% of cells showed a cytopathic effect (CPE). Cells in the medium were freeze-thawed once, centrifuged at 12,000 rpm for 10 min, and the resulting supernatant was aliquoted and stored at −80°C. Virus titers were determined by TCID₅₀ assay as described below.

Incubation of GETV with soluble GAGs.

GETV virions at a multiplicity of infection (MOI) 0.1 were pretreated with 1 mg/mL heparin (Hp) (Solarbio) or 1 mg/mL chondroitin sulfate (CS) (Solarbio) at 37°C for 30 min before inoculation of monolayer cells. BSA-treated (1 mg/mL) virions served as control. The inoculum was removed after incubation at 37°C for 1 h, and cells were washed 3 times with phosphate buffer saline (PBS) and incubated at 37°C for 24 h in a maintenance medium. Infectivity was then quantified using qRT-PCR and virus titers.

Incubation of GETV with soluble heparin.

GETV virions (MOI of 0.1) were incubated with Hp (0, 0.125, 0.25, and 0.50 mg/mL) for 30 min prior to attachment to monolayer cells. The inoculum was removed after incubation at 37°C for 1 h, and cells were washed and incubated at 37°C for 24 h in a maintenance medium. Infectivity was quantified using indirect immunofluorescence and virus titer assay.

Virus replication assay.

Cells monolayers were incubated with the virus inoculum (MOI = 0.1) at 37°C for 1 h. Afterward, the inoculum was discarded, cells were washed 3 times with PBS and cultured for 24 h. The supernatant was collected for virus titer assay by TCID₅₀, and the cells were lysed for qRT-PCR or Western blot.

Virus attachment assay.

Cells were seeded onto a 24-well plate and cultured until a confluent monolayer was formed. Cells were then pre-cooled to 4°C for 10 min prior to incubation with the pre-chilled virus inoculum (MOI = 1) at 4°C for 1 h to allow the virus to attach to the cell surface. The following operations were also performed on ice: The inoculum was removed, cell monolayers were washed 3 times with pre-cooled PBS to remove unbound virus, and TRIzol (500 μL) was added to lyse the cells for RNA isolation. Subsequently, qRT-PCR was used to determine virus genome copy numbers by detecting E2 gene and normalized to untreated controls.

Plaque assays for detecting bound virus.

Cells monolayers were pre-cooled to 4°C for 10 min prior to incubation with the virus (MOI = 0.01) for 1 h at 4°C. Cells were washed 3 times with pre-cooled PBS and then covered with a mixture (1:1) of methylcellulose and DMEM containing 1% penicillin-streptomycin and 2% FBS. Cells were then cultivated at 37°C for 48 h. The methylcellulose overlay was removed, cells were washed 3 times with PBS, and stained with a mixture of crystal violet and ammonium oxalate (1:4) for 1 h. PBS was used to wash away the excess crystal violet, and the number of plaques in each well was counted after standing to dry. Theoretically, each plaque represents an infectious virus particle that adsorbed during the incubation period to a cell.

Preincubation of cells with sodium chlorate.

As previously described (31, 46), BHK-21 and PK-15 cells were passaged twice in a culture medium containing 10, 25, or 50 mM sodium chlorate (NaClO₃). Some cell dishes were subsequently cultured in the presence of 0.8 mM sodium sulfate (Na₂SO₄) or in a medium without NaClO₃ for 48 h to reverse the effect of NaClO₃ on cells with two methods. Some cell dishes were treated with NaClO₃ continuously. The untreated cells were also used as a control. The culture medium was removed, and the cells were then incubated with GETV (MOI = 0.1) for 1 h and washed 3 times with PBS. Virus replication was determined after being cultivated at 37°C for 24 h using qRT-PCR, Western blot, and virus titer. After virus (MOI = 0.01) attachment at 4°C, cultured at 37°C for 48 h until plaques formation. The number of plaques at different NaClO₃ concentrations was determined.

Incubation of viruses with sodium chloride.

Viruses were diluted into DMEM supplemented with increasing sodium chloride (NaCl) concentrations (0 mM, 150 mM, 300 mM, final concentration) prior to incubation of cells for 1 h. Cells were then cultured for 24 h. Subsequently, the infection efficiency was determined by qRT-PCR, virus titer assay, and counting of plaques. Cells were lysed directly with TRIzol followed by qRT-PCR to detect changes in viral adsorption capacity.

Heparin-conjugated magnetic beads binding assay.

An equal amount of heparin- and His-conjugated magnetic beads (BEAVER) were washed twice with binding buffer (50 mM Tris-HCl, pH 8.0). The same number of infectious virions (10⁶/TCID₅₀) of rGETV-HN and E2-K253R were diluted to 100 μL with binding buffer, and 70 μL were incubated with magnetic beads (30 μL) using a vertical mixer for 1 h at 4°C. The remaining 30 μL virus were used as input samples. An equal volume of DMEM is diluted as a control. After magnetic separation of beads and supernatant, the latter was removed, and beads were resuspended by vortexing for 1 min in binding buffer, then magnetically separated, and the supernatant was removed. This washing step was repeated 3 times. Then 50 μL elution buffer (50 mM Tris-HCl, 2 M NaCl, pH 8.0) was added, samples vortexed for 1 min, and incubated for 15 min in a vertical mixer at room temperature. After magnetic separation, the supernatant was collected. Then, 10 μL of DMEM control, input virus, and elution samples were used for Western blot with polyclonal antiserum (prepared in our laboratory, 1:500) against GETV-cap (capsid). The density of protein bands in the blot was determined and the percentage of virus bound to beads was calculated as the mean of bound virus divided by the total input virus as previously described (30).

Mouse virus infections and pathogenesis studies.

In the experiment comparing the virulence of rGETV-HN and E2-K253R, 2-day-old ICR mice were mock-infected or infected with viruses (25 μL of 10⁶/TCID₅₀ in DMEM) either i.c. or s.c. injected. Virus- and mock-infected mice were observed daily for 14 days and weighed at 12-h intervals until symptoms or death appeared. Average survival times (ASTs) and percent mortality were calculated.

Mice were euthanized at 24 h and 48 h after being injected to determine virus titers in tissues. Lung, kidney, trachea, turbinate, knee joint, and brain tissues were collected, and the net weight of the tissues was determined. PBS was added, samples were homogenized and freeze-thawing once. After centrifugation at 12,000 rpm for 5 min, supernatants were aspirated for virus titer assay.

Virus clearance in vivo.

Four weight-matched male 5-week-old ICR mice were anesthetized intraperitoneally with tribromoethanol (0.5 mg per g). Mice were inoculated into the tail vein with 10⁶/TCID₅₀ of virus in 25 μL DMEM. Serum samples were collected from the retro-orbital plexus at 5-, 10-, 20-, 40-, and 60-min postinjection (p.i.) using a 20 μL capillary tube. Blood was allowed to clot at room temperature for 1 h, centrifuged for 10 min (3,000 rpm) at 4°C, and supernatants were collected for storage at −80°C. Determination of virus content in serum was performed by TCID₅₀ in Vero cells.

Cell viability assay.

The cytotoxic effect of Hp, CS, and NaCl on BHK-21 and PK-15 cells was assessed using the Cell Counting Kit-8 (CCK-8, APE×BIO). Cells monolayers grown in 96-well plates were incubated in a medium with increasing Hp, CS, or NaCl concentrations and continued cultivating for 2 h at 37°C. The medium was then replaced with fresh DMEM containing 2% FBS, and cells were cultured for 24 h at 37°C. After 24 h, the cells were washed with PBS and incubated with 90 μL DMEM and 10 μL of CCK-8 solution at 37°C for 2 h. Absorbance was measured with a microplate reader (Tecan, M200 PRO, CH) at 450 nm. Cytotoxicity was expressed according to the following formula:

$$\text{Cytotoxicity\hspace{0.17em}}\left(\%\right)=\left(\text{Abs\hspace{0.17em}sample}\right)-\left(\text{Abs\hspace{0.17em}blank}\right)\left(\text{Abs\hspace{0.17em}negative\hspace{0.17em}control}\right)-\left(\text{Abs\hspace{0.17em}blank}\right)\times \text{\hspace{0.17em}1}00$$

Quantitative real-time PCR (qRT-PCR).

According to the manufacturer’s instructions, viral RNA was extracted from cells using TRIzol reagent (Vazyme), and cDNA was synthesized with a Hiscript II 1st Strand cDNA Synthesis Kit (Vazyme). RNA (1 μg) was used for reverse transcription with Oligo-(dT) primer. Equal volumes of cDNA were used for qRT-PCR by SYBR (Vazyme) using a Light Cycler 96 real-time PCR system (Roche Diagnostics). Gene expression assessment was done in triplicate using sets of primers reactive with the GETV E2 gene, β-actin or GAPDH-human was used as an internal control (Table 1). The data were analyzed using the cycle threshold (2^−△△Ct) method (47).

Western blot.

Cells were lysed in RIPA buffer (Beyotime), an appropriate amount of SDS-PAGE loading buffer (Solarbio) was added, and samples were boiled for 10 min. The same sample volume (10 μL) was subjected to SDS-PAGE and then transferred to a nitrocellulose membrane (NC, GE Amersham Biosciences). The membranes were blocked with 2% (wt/vol) nonfat milk for 1 h, followed by primary antibody incubation for 1 h (anti-GETV-E2 (polyclonal antiserum prepared in our laboratory, 1:500), anti-GETV-cap polyclonal Abs (pAbs), or anti-β-actin (Abmart, 1:1,000)), further incubated for 1 h with secondary antibody (HRP-anti-rabbit IgG or HRP-anti-mouse IgG) (KPL, 1:10,000) at room temperature. Before and after incubating the antibodies, membranes were washed 5 times with PBST (0.1% Tween 20). Protein bands were visualized with an LI-COOdyssey infrared image scanner.

Immunofluorescence assay.

After being infected for 24 h, cells were fixed with 4% paraformaldehyde at room temperature for 1 h, permeabilized with 0.1% Triton X-100 in PBS for 30 min, and washed 3 times. Cells were blocked with 5% nonfat milk for 1 h and incubated with the anti-GETV-E2 pAbs for 1 h at 37°C. Cells were washed with PBS 3 times before incubation with FITC-conjugated secondary antibody (KPL, 1:1,000) for 1 h at 37°C. After washing with PBS 3 times, cells were incubated with 4, 6-diamidino-2-phenylindole (DAPI, Solarbio, 1:100) for 10 min prior to observing using a differential fluorescence microscope (Nikon), and two-color fluorescence images were recorded from at least three separate experiments.

Virus titration by TCID₅₀.

Cells grown in 96-well plates were infected with 10-fold serially diluted virus inoculum and incubated at 37°C for 2 days until obvious CPE developed. Each dilution had six replicates, and each sample was titrated in triplicate. The TCID₅₀ was calculated using the Reed and Munch method (48).

Statistical analyses.

Statistical analysis was performed using the Student's t test using the Prism 8.0 software (GraphPad Software). Error bars, P values, and numbers of biological replicates (n) are specified in the relevant figure legends.

Animal experiments.

The animal experiments were approved by the Institutional Animal Care and Use Committee of Nanjing Agricultural University, Nanjing, China (no. SYXK2017-0007; February 2017).

Data availability.

Sequences for the GETV strains are available in GenBank as follows: GETV-GX, MZ736796; GETV-HN, MZ736801; and GETV-FJ, MZ736799.