Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jun 21.
Published in final edited form as: Sci Transl Med. 2026 Apr 8;18(844):eaec2923. doi: 10.1126/scitranslmed.aec2923

Design of a neutralizing and protective pan-encephalitic alphavirus receptor decoy protein Overline: INFECTIOUS DISEASES

Sathvik Palakurty 1,2,*, Saravanan Raju 1,7,*, Alan Sariol 2, Ngan Wagoner 2, Theron C Gilliland Jr 3, William B Klimstra 3, Michael S Diamond 1,2,4,5,6
PMCID: PMC13282995  NIHMSID: NIHMS2181888  PMID: 41950307

Abstract

Venezuelan (VEEV), Eastern (EEEV), and Western (WEEV) equine encephalitis viruses are alphaviruses from different serocomplexes that cause neurological disease in humans. Given their antigenic distance, it has been challenging to isolate cross-reactive antibodies that neutralize infection of multiple medically relevant encephalitic alphaviruses. Recently, distinct entry receptors were identified for these encephalitic alphaviruses: LDLRAD3 for VEEV, VLDLR for EEEV and some strains of WEEV, and PCDH10 for WEEV. Here, using structure-guided mutagenesis, we generated a soluble chimeric protein derived from the LA1 domain of LDLRAD3 and the LA2 domain of VLDLR, termed LDLRAD3-vLA1-VLDLR-LA2, which neutralized infection of VEEV, EEEV, and some WEEV strains in cell culture and protected mice from infection. Structural analysis of this engineered decoy revealed binding to distinct sites on each virus, which corresponded to those engaged by their endogenous entry receptors. We extended the neutralizing and protective capacity to contemporary WEEV strains by adding a single extracellular cadherin domain from avian PCDH10 through a stabilized symmetric bispecific decoy scaffold. Our designed receptor decoy serves as a possible countermeasure against multiple encephalitic alphaviruses, and this design platform could be harnassed to develop therapeutic agents against viruses from other families.

One sentence summary

An engineered protein receptor decoy protects mice from lethal infection caused by three antigenically distinct encephalitic alphaviruses.

Editor’s Summary:

Identifying one neutralizing treatment against multiple encephalitic alphaviruses would offer broad protection against associated neurological disease. However, this has been challenging to develop due to phylogenetic differences between viruses. Here, Palakurty et al. utilized the structure of unique entry receptors for Venezuelan (VEEV), Eastern (EEEV), and Western (WEEV) equine encephalitis viruses, followed by structure-guided mutagenesis to develop a soluble chimeric neutralizing decoy protein. The LA1 domain of LDLRAD3 (VEEV), the LA2 domain of VLDLR (EEEV and some WEEV strains), and an extracellular cadherin domain from avian PCDH10 (WEEV) were combined to generate a multi-potent decoy protein which conferred broad protection against infection in vitro and in vivo. This approach not just offers one uniform treatment against VEEV, EEV and WEEV, but also offers a platform approach for designing novel, uniform neutralizing treatments against other viral families.

INTRODUCTION

In the early 20th century, alphaviruses were isolated from horses with acute neurological disease and named Western (WEEV) and Eastern equine encephalitis (EEEV) viruses, depending on the geographic location in the United States1,2. In 1938, encephalitis was also detected in a Venezuelan child which was caused by a different neurotropic alphavirus, namely Venezuelan equine encephalitis virus (VEEV)3. The proteins of VEEV, EEEV, and WEEV are encoded by approximately 11.7 kilobases in a positive sense, single-stranded RNA genome with two open reading frames (ORFs). The first ORF encodes four nonstructural proteins responsible for replication and immune evasion. The second ORF encodes a structural polyprotein that packages the RNA genome into virions and mediates entry into target cells. The structural glycoproteins of VEEV, EEEV, and WEEV share only about 50% pairwise sequence identity4 with most conserved regions of the glycoproteins not readily accessible on the virion surface. Consequently, cross-neutralizing antibody responses against VEEV, EEEV, and WEEV are rare following experimental infection or vaccination with a single encephalitic alphavirus species5.

Several proteins containing LDLR class A (LA) domains were recently described as entry receptors for alphaviruses. Entry of VEEV is mediated by LDLRAD36, whereas EEEV, Semliki Forest virus (SFV), and some strains of WEEV can engage VLDLR and LRP87–9. LDLR family members also have been reported as candidate entry receptors of viruses from other genera: LRP1 for Rift valley fever (RVFV)10, Oropouche (OROV)11, Dengue (DENV)12, and yellow fever (YFV)13 viruses; LDLR for human rhinovirus 2 (HRV2), vesicular stomatitis, Crimean Congo hemorrhagic fever, West Nile (WNV), and Japanese encephalitis (JEV) viruses14–20; LRP8 for tick-borne encephalitis virus and YFV 13,21,22; and multiple LDLR family members for YFV13. We reasoned that combination of individual LA domains from different LDLR family members might provide a means for developing a soluble receptor decoy molecule with broader inhibitory activity. LA domains are approximately 40 amino acids in length with conserved cysteine residues required for folding and a negatively charged binding pocket containing conserved acidic residues that bind a calcium ion23,24. Apart from these conserved residues that define the LA domain, the sequences of LA domains are highly variable, which is thought to contribute to their specificity for endogenous ligands and possibly, to viral glycoproteins.

Structural and functional studies of LA domain containing receptors and encephalitic alphaviruses revealed that their modes of engagement can vary8,25–31. LDLRAD3-LA1 binds in a cleft formed by E1-E2 heterodimers of VEEV6,30,31. Analogously, VLDLR-LA1 binds in a cleft formed by E1-E2 heterodimers of EEEV and WEEV28,29,32. Whereas an extensive interface is formed between LDLRAD3-LA1 and VEEV, VLDLR-LA1 engages a smaller interface to bind EEEV and WEEV, and requires additional interactions between VLDLR-LA2 and a site outside of the E1-E2 cleft in the A or B domain of E2. In addition to LA domain containing proteins, WEEV uniquely engages protocadherin 10 (PCDH10) in the E1-E2 cleft with residues overlapping those engaged by VLDLR-LA125,26,33

Despite the structural advances in defining alphavirus-receptor interactions, the molecular basis of the differential specificity of LA domain interactions for VEEV, EEEV, or WEEV remains unclear. Given the relatively small surface area engaged by VLDLR-LA1 in the cleft of EEEV and WEEV, we reasoned that mutation of even a few residues might enable a modified LDLRAD3- LA1 variant to bind to both EEEV and WEEV. Here, we engineered variant LDLRAD3-LA1 molecules linked in tandem to an unmutated VLDLR-LA2 domain. These proteins, along with the addition of a domain of PCDH10, were incorporated into different antibody-like scaffolds and tested for panencephalitic alphavirus activity against VEEV, EEEV, and WEEV in cell culture and in mice. Our study goal was to use functional and structural information of virus-receptor interactions to design and generate a single molecule that could inhibit infection of VEEV, EEEV, and WEEV in cell culture and in animals.

RESULTS

A single mutation in LDLRAD3-LA1 enables binding to EEEV virus-like particles.

EEEV, VEEV, and WEEV show poor conservation of the surface-exposed epitopes within the E1 and E2 glycoproteins (fig. S1). Yet, all three viral species are competent to bind distinct LA domain(s) which could be leveraged to design a pan-encephalitic decoy molecule. LDLRAD3-LA1 and VLDLR-LA1 bind VEEV and EEEV, respectively, in a cleft formed by E2-E1 heterodimers (Fig. 1A) 28–32. LDLRAD3-LA1 interacts with VEEV through extensive contacts in the cleft 30,31. At one interface, a hydrophobic patch of VEEV E2 engages residues G33, P32 and W47 of LDLRAD3-LA1. More distally along the same interface, electrostatic interactions form between VEEV E2-R64 and residues D50 and D56 of LDLRAD3-LA1 (Fig. 1A). The buried surface area of LDLRAD3-LA1 and VEEV is approximately 900 Å2. For EEEV, the buried surface area between VLDLR-LA1 and the E2 β-connector is relatively small (approximately 300 Å2). Binding of VLDLR-LA1 to EEEV E2 is mediated by residues E2-H155, E2-K156, E2-R157 of EEEV E2, which form electrostatic interactions with VLDLR-LA1 D53, D55, D57 and hydrophobic and π-cation interactions with W50.

Figure 1. Mutation of a single residue enables LDLRAD3-LA1 binding to EEEV.

Figure 1.

Open in a new tab

(A) (Left) Model of VEEV complexed with LDLRAD3 with inset showing the interface of VEEV and LDLRAD3 L52 (PDB: 7FFL31). (Right) Model of EEEV complexed with VLDLR with inset showing the interface of EEEV and VLDLR at the E1-E2 ‘cleft’ site (PDB: 8UFC29). (B) Alignment of LDLRAD3-LA1 and VLDLR-LA1 with interface contacts labeled. Identical (red) or biochemically similar (white) residues are boxed. Mutated residues in the engineered decoy are indicated with an arrow above the alignment. The added serine present in the engineered decoy is indicated in purple. (C) Surface representation of LDLRAD3-LA1 with regions that are identical between VLDLR-LA1 and LDLRAD3-LA1 colored in blue and differing regions colored in yellow. (D) Biolayer interferometry (BLI) sensorgrams showing binding of VEEV TC-83 and EEEV PE6 virus-like particles (VLPs) to biosensors loaded with LDLRAD3-LA1-Fc, LDLRAD3-LA1(L52D)-Fc, VLDLR-LA1-Fc, or mouse MXRA8-Fc. Representative data from 3 independent experiments is shown.

We hypothesized that LDLRAD3-LA1 could be engineered to bind to EEEV with few amino acid substitutions to create similar electrostatic interactions between the mutated LDLRAD3-LA1 and EEEV E2 as seen for VLDLR LA1 and EEEV E2 (Fig. 1B) 28,29,32. LDLRAD3-LA1 and VLDLR-LA1 share only about 50% sequence identity with many surface exposed residues differing between the two proteins (Fig. 1B and C). Given our prior structural and mutagenesis data28,29,32, we focused on residue D55 of VLDLR-LA1, which forms a requisite charge contact with EEEV E2-R157 in the cleft formed by E2-E1 heterodimers. Because an acidic residue in VLDLR-LA1 aligns with a hydrophobic leucine (L52) residue in LDLRAD3-LA1 (Fig. 1B and C), we hypothesized that mutation of LDLRAD3-LA1 L52 to the corresponding aspartic acid residue might enable EEEV binding. To test this idea, we expressed recombinant receptors as Fc-fusion proteins and assessed their binding to virus-like particles (VLPs, encoded by C-E3-E2-E1). Biolayer interferometry (BLI) sensors coated with LDLRAD3-LA1 [L52D]-Fc captured EEEV FL93-939 and VEEV TC-83 VLPs, whereas sensors coated with wild-type LDLRAD3-LA1-Fc bound only to VEEV TC-83 VLPs (Fig. 1D and fig. S2A). However, as expected, LDLRAD3-LA1 [L52D]-Fc did not bind to VLPs of a more distantly related arthritogenic alphavirus, chikungunya virus (CHIKV) (fig. S2B). Thus, a single amino acid substitution in LDLRAD3-LA1 expands the binding breadth to both VEEV and EEEV.

To determine if a LDLRAD3-LA1-Fc soluble decoy receptor with the L52D substitution was sufficient to inhibit EEEV and VEEV infection, we performed neutralization experiments with Sindbis virus (SINV) chimera expressing the non-structural genes of SINV and structural genes of EEEV (SINV-EEEV) or VEEV (SINV-VEEV) along with a cotranslational GFP reporter to monitor infection (fig. S2C).

Mutations to LDLRAD3-LA1 and addition of VLDLR-LA2 enhance neutralization of EEEV.

Although LDLRAD3-LA1 [L52D]-Fc retained the ability to neutralize SINV-VEEV, it did not inhibit SINV-EEEV infection (Fig. 2A and B). The poor neutralization of EEEV by LDLRAD3-LA1 [L52D]-Fc might have been anticipated since inhibition of EEEV by VLDLR-Fc decoys requires binding of at least two LA domains to distinct sites29. For EEEV, LA1 binds to the E1-E2 cleft, and LA2 binds the A domain of E2 (Fig. 1A)28,29,32. To account for the two-domain binding pattern required for neutralization, we generated LDLRAD3-LA1-VLDLR-LA2-Fc and LDLRAD3-LA1 [L52D]-VLDLR-LA2-Fc proteins (fig. S3A). Whereas LDLRAD3-LA1-VLDLR-LA2-Fc did not neutralize SINV-EEEV at 10 μg/ml, LDLRAD3-LA1 [L52D]-VLDLR-LA2-Fc did, but only weakly (Fig. 2A and C). Thus, more substitutions were needed for efficient neutralization of SINV-EEEV.

Figure 2. A chimeric LDLRAD3-vLA1-VLDLR-LA2-Fc variant neutralizes encephalitic alphaviruses in vitro.

Figure 2.

Open in a new tab

(A) Diagram of LDLRAD3-LA1-Fc and LDLRAD3-LA1-VLDLR-LA2-Fc. (B and C) SINV-EEEV-GFP and SINV-VEEV-GFP infection in HEK293T cells following incubation with indicated one LA domain (B) or two LA domain (C) soluble receptor proteins at 10 μg/ml. Infection was assessed by the frequency of GFP+ cells using flow cytometry. (D and E) Dose response analysis (D) and EC50 values (E) of indicated soluble decoy inhibitors against SINV-EEEV-GFP and SINV-VEEV-GFP infection in HEK293T cells. Infection was assessed by the frequency of GFP+ cells using flow cytometry. Error bars indicate standard error of the means (SEM). Data are normalized to infection with no decoy receptor. n.m. indicates an EC50 value that was not measurable. Data in (B to E) were pooled from 3 independent experiments, each with two technical replicates. Statistical analysis in (B) and (C) was performed using a one-way ANOVA with Dunnett’s post-test; n.s., not significant; ** p < 0.01.

To improve the neutralizing activity of LDLRAD3-LA1 [L52D]-VLDLR-LA2-Fc, we focused on the interface between VLDLR-LA1 domain and EEEV, and identified additional residues differing between VLDLR-LA1 and LDLRAD3-LA1 (Fig. 1B). The region of LDLRAD3-LA1 contiguous with residue L52 corresponds to a more negatively charged site of VLDLR-LA1 that forms contacts with the positively charged EEEV residue E2-K156. To account for this charge interaction, we introduced a P53E substitution into LDLRAD3-LA1 [L52D] to match the residue in VLDLR-LA1. We also introduced an N34Q substitution into a region that corresponds to a face of VLDLR-LA1 that forms contacts with the E1 fusion loop and positively charged E2-H175 and E2-K221 residues of EEEV. The resulting LDLRAD3-LA1 [N34Q, L52D, P53E]-VLDLR-LA2-Fc decoy protein fully neutralized infection by SINV-EEEV at 10 μg/ml (Fig. 2C). However, a dose response curve analysis showed that LDLRAD3-LA1 [N34Q, L52D, P53E]-VLDLR-LA2-Fc was less potent than VLDLR-LA1-LA2-Fc at neutralizing SINV-EEEV (half-maximal effective inhibitory concentration (EC50) values of 1,847 ng/ml and 364 ng/ml, respectively) (Fig. 2D and E). To further improve the activity of LDLRAD3-LA1 [N34Q, L52D, P53E]-VLDLR-LA2-Fc molecule, we tested additional mutations: E28K, M36Q, or K58G. These substitutions were chosen to replace residues in LDLRAD3-LA1which do not interface with VEEV and correspond to VLDLR-LA1,conseqeuntly promoting the interaction strength with EEEV30,31. Since LDLRAD3-LA1 [N34Q, L52D, P53E]-VLDLR-LA2-Fc fully neutralized SINV-EEEV at 10 μg/ml, we tested these mutant proteins at a lower dose, 1 μg/ml. The addition of K58G, which created LDLRAD3-LA1 [N34Q, L52D, P53E, K58G]-VLDLR-LA2-Fc enabled neutralization of SINV-EEEV at 1 μg/ml (fig. S3B).

Since avid EEEV binding to VLDLR requires concurrent binding to more than one LA domain site29, we reasoned that the linker between LA1 and LA2 should accommodate engagement of both domains. Accordingly, we next focused on the C-terminal region of the LDLRAD3-LA1. VLDLR-LA1 (residues 31-69) contains two additional residues (V68 and K69) after the terminal cysteine, whereas LDLRAD3-LA1 (residues 28-65) contains only a proline residue (P65). We replaced the terminal P65 residue of LDLRAD3-LA1 [N34Q, L52D, P53E] with V65 and K66, which lengthens the LA1 domain by a single amino acid. However, the replacement of the terminal P65 with V65 and K66 did not improve neutralization activity (fig. S3B). We next tested a separate substitution, in which we added a terminal serine, ins66S, to allow for more flexibility between the two LA domains. The ins66S mutant improved neutralization of EEEV when compared to the parental decoy sequence (fig. S3B).

We then combined candidate mutations and insertions for testing of neutralization. LDLRAD3-LA1 [N34Q, L52D, P53E, K58G, ins66S]-VLDLR-LA2-Fc (hereafter, LDLRAD3-vLA1-VLDLR-LA2-Fc) had potent inhibitory activity of SINV-EEEV (EC50 of 237 ng/ml; 95% CI of 160 to 350) (Fig. 2D and E). LDLRAD3-vLA1-VLDLR-LA2-Fc showed greater neutralizing activity (approximately 300-fold) against SINV-VEEV than the parental LDLRAD3-LA1-Fc. In comparison, LDLRAD3-LBD-Fc, which consists of 3 LA domains (LA1-LA2-LA3), neutralized SINV-VEEV less effectively (EC50 of 6,487 ng/ml; 95% CI 5,012 to 8,483), suggesting that length of the decoy molecule alone does not explain the enhanced inhibitory activity of LDLRAD3-vLA1-VLDLR-LA2-Fc (Fig. 2D and E).

The introduction of L52D and N34Q increased neutralization potency against VEEV, but the addition of P53E had little effect. The further addition of K58G, which is not seen at either interface of LDLRAD3 and VEEV, increased neutralization potency against VEEV in the context of N34Q, L52D, and P53E substitutions (fig. S3C). Our mutational progression makes the electrostatic potential of the buried surface interacting with VEEV E2-R64 more negative, more closely resembling that of VLDLR-LA1 (fig. S3D).

To confirm the geometry of engagement, we performed cryo-electron microscopy analysis of monomeric forms of LDLRAD3-vLA1-VLDLR-LA2 complexed with VEEV and EEEV VLPs. VEEV VLP complexed with LDLRAD3-vLA1-VLDLR-LA2 shows density in the cleft formed between E1-E2 heterodimers. For the EEEV VLP complex with LDLRAD3-vLA1-VLDLR-LA2, we observed two densities roughly the size of an LA domain at the E1-E2 cleft site and in the E2-A domain (fig. S4 and S5), as seen with EEEV VLPs bound to VLDLR-LA1-LA228,29,32. Thus, LDLRAD3-vLA1-VLDLR-LA2 engages different sites on alphavirus virions corresponding to those bound by LDLRAD3 and VEEV and VLDLR and EEEV. An examination of the interface of LDLRAD3-vLA1 and EEEV VLPs showed that the introduced L52D mutation interacts with EEEV E2-K156 and is stabilized by the LDLRAD3-vLA1 W47 residue, similar to interface chemistry between EEEV and VLDLR-LA1 (fig. S6A). Furthermore, the introduced L52D substitution interacts with VEEV E2-R64, which also may explain the enhanced neutralization potency of the mutant LDLRAD3-vLA1 proteins (fig. S6B).

LDLRAD3-vLA1-VLDLR-LA2-Fc efficiently neutralizes WEEV strains that engage VLDLR.

Most WEEV strains that harbor E2-K181 or E2-K81 residues can engage VLDLR25,26,33. We and others have shown that VLDLR-LA(1–2)-Fc proteins can neutralize WEEV McMillan (encodes E2-K181) and WEEV Fleming (encodes E2-K81) but not WEEV CBA87 (encdes E2-E81 and E2-E181)25,26,33. Although, compared to EEEV, VLDLR-LA(1–2) engages WEEV through a distinct set of residues on the viral glycoprotein within the E1-E2 cleft fig. S7A), both viruses engage the same residues of VLDLR-LA1 (including W50, D55. and E56; Fig. 1B). Accordingly, we next tested whether our one or two LA domain engineered decoy molecules could neutralize WEEV strains. Similar to that seen with EEEV25, two LA domains were required for neutralization of WEEV (Fig. 3A, fig. S7B). The substitutions that enhanced neutralization of EEEV functioned similarly against WEEV McMillan and Fleming strains. However, and as expected, WEEV CBA87 was not neutralized by any LDLRAD3-VLDLR-LA domain containing molecule (Fig. 3A). Thus, LDLRAD3-vLA1-VLDLR-LA2-Fc can neutralize VEEV, EEEV, and VLDLR-engaging strains of WEEV demonstrating near pan-encephalitic inhibitory activity in cell culture.

Figure 3. Incorporation of PCDH10 enables generation of a decoy molecule that neutralizes additional WEEV strains in vitro.

Figure 3.

Open in a new tab

(A) SINV-WEEV-GFP McMillan, Fleming, and CBA87 infection in HEK293T cells following incubation with soluble receptor decoy proteins at 10 μg/ml. Infection was analyzed by the frequency of GFP+ cells using flow cytometry. Data are pooled from 3 independent experiments. (B) Diagram of bispecific topologies. Knob-in-hole topology is generated by heterodimerization of PCDH10-EC1 and LDLRAD3-vLA1-VLDLR-LA3. Daisy Chain A: PCDH10-EC1 at the N-terminus of LDLRAD3-vLA1-VLDLR-LA2-Fc; Daisy Chain B: LDLRAD3-vLA1-VLDLR-LA2 at the N-terminus of PCDH10-EC1-Fc; Daisy Chain C: PCDH10-EC1 at the C-terminus of LDLRAD3-vLA1-VLDLR-LA2-Fc. Heavy Chain : Light Chain topology: PCDH10-EC1 is appended to the N-terminus of CΚ, and LDLRAD3-vLA1-VLDLR-LA2 is appended to the N-terminus of CH1–CH2–CH3. (C) Dose response analysis of indicated bispecific decoy inhibitors against SINV-EEEV FL93, SINV-VEEV TrD, SINV-WEEV McMillan, and SINV-WEEV CBA87. Data are pooled from 2 to 3 independent experiments; error bars indicate standard error of the means (SEM). (D) Structural model (PDB:1HZH55) showing the placement of stabilizing mutations at the CH1–Cκ interface and the AlphaFold3 (AF3)-predicted model with engineered mutations. Surface view of model (left) and close-up view of CH1–Cκ interface (center-left) with Cκ displayed as a ribbon. Detailed view of interface (center-right) and model of AF3-predicted interface (right). Statistical analysis in (A) was performed using a one-way ANOVA with Dunnett’s post-test; n.s., not significant; * p < 0.05.

LDLRAD3-vLA1-VLDLR-LA2-Fc protein protects mice from infection with EEEV, VEEV and some strains of WEEV.

We next tested the ability of LDLRAD3-vLA1-VLDLR-LA2-Fc to protect against in vivo challenges by VEEV, EEEV, and WEEV. Six-week-old mice were treated with LDLRAD3-vLA1-VLDLR-LA2-Fc or negative control MXRA8-Fc proteins 6 h prior to subcutaneous challenge with VEEV (Trinidad Donkey [TrD] strain), EEEV (FL93-939 strain), or WEEV (McMillan strain). In contrast to the MXRA8-Fc control, LDLRAD3-vLA1-VLDLR-LA2-Fc conferred near complete protection against mortality caused by all three encephalitic alphaviruses (fig. S7C). However, prophylaxis of mice with LDLRAD3-vLA1-VLDLR-LA2-Fc failed to protect against a WEEV strain expressing the structural genes of the CBA87 strain, which does not engage VLDLR (fig. S7C)7,8,34.

A symmetric, bispecific topology incorporates PCDH10-EC1 to neutralize additional WEEV strains.

Most group B and C WEEV strains do not utilize LA domain containing proteins (e.g., VLDLR or LDLR) as entry receptors7,8. However, all WEEV strains tested can use the first extracellular cadherin repeat (EC1) of avian PCDH10 for entry, whereas VEEV and EEEV do not engage this receptor8,25. Using this knowledge, we sought to design a more complete pan-encephalitic decoy with activity against all WEEV strains. We undertook several approaches to incorporate PCDH10-EC1 into the LDLRAD3-vLA1-VLDLR-LA2-Fc protein (Fig. 3B). We first generated chimeric molecules that placed PCDH10-EC1 at the N-terminus of LDLRAD3-vLA1-VLDLR-LA2. However, this molecule lacked potent neutralizing activity against EEEV (Fig. 3C). When we positioned PCDH10-EC1 in between the LDLRAD3-vLA1-VLDLR-LA2 domain and the Fc domain, the resulting molecule poorly neutralized WEEV CBA87 infection, yet maintained inhibitory activity against EEEV and VEEV. We next generated a bispecific Fc-fusion using knob-in-hole technology35 (Fig. 3B and C, fig. S8A), such that the bispecific molecule is formed by heterodimerization of LDLRAD3-vLA1-VLDLR-LA2 Fc monomer and PCDH10-EC1 Fc monomer. However, this molecule poorly neutralized SINV-EEEV infection (EC50 > 10 μg/ml), although some inhibitory activity was detected at high concentrations (Fig. 3C, fig. S8B). To attain bivalency of each receptor component, we designed molecules with a symmetric bispecific topology. Separated by a 27-amino acid Gly-Ser linker, PCDH10-EC1 was added to the N-terminus of LDLRAD3-vLA1-VLDLR-LA2-Fc, which we denoted as Daisy Chain A (Fig. 3B). Reciprocally, again with a 27-amino acid Gly-Ser linker, we added LDLRAD3-vLA1-VLDLR-LA2 to the N-terminus of PCDH10-EC1-Fc, denoted as Daisy Chain B. Finally, we added the LDLRAD3-vLA1-VLDLR-LA2 decoy to the C-terminus of PCDH10-EC1-Fc with a 23-amino acid Gly-Ser linker, denoted as Daisy Chain C. All three proteins failed to neutralize EEEV, VEEV, and both WEEV strains efficiently (Fig. 3C, fig. S8A).

Instead of modifying linker length and flexibility, we pursued a different design strategy by engineering a bivalent decoy molecule based on the molecular arrangement of an antibody (Fig. 3B). We exchanged the variable fragment (VH) of the human IgG1 heavy chain with LDLRAD3-vLA1-VLDLR-LA2 and then fused PCDH10-EC1 to the N-terminus of the constant region of the kappa light (Cκ), with the idea that co-expression would produce a bivalent molecule through heterodimerization of CH1 and Ck. However, this molecule showed incomplete heterodimerization, possibly due to lack of intramolecular contacts between variable regions that were replaced with receptor ectodomains36. To stabilize heterodimerization of the molecule, we introduced cysteine37 and hydrophobic38 residues (F9C and S66V in CH1; S14C and S69L in Ck) (Fig. 3D). Individually, each set of stabilizing mutations increased dimerization to a small extent. However, combining both sets of substitutions resulted in efficient heterodimerization and protein expression (Fig. 3B fig. S8C). The resulting molecule, PCDH10-EC1:LDLRAD3-vLA1-VLDLR-LA2-Fc, neutralized all WEEV strains tested and retained potent inhibitory activity against EEEV and VEEV in cell culture (Fig. 3C).

PCDH10-EC1:LDLRAD3-vLA1-VLDLR-LA2-Fc protects mice from EEEV, VEEV, and WEEV infection.

We tested the inhibitory of PCDH10-EC1:LDLRAD3 vLA1-VLDLR-LA2-Fc in vivo. Mice were administered a single 100 μg dose (approximately 4 mg/kg) by intraperitoneal injection at 24 h prior to subcutaneous challenge with EEEV (Madariaga [MADV] strain), WEEV (McMillan or CBA87 strains), or VEEV (ZPC-738 or TrD strain). Compared to the respective control proteins (LDLRAD3-LA1-Fc for WEEV and EEEV; PCDH10-EC1-Fc for VEEV), mice treated with PCDH10-EC1:LDLRAD3-vLA1-VLDLR-LA2-Fc showed higher survival rates after challenge with all encephalitic alphaviruses tested (Fig. 4A, fig. S8D). Consistent with this result, mice treated with PCDH10-EC1:LDLRAD3-vLA1-VLDLR-LA2-Fc also had substantially lower amounts of VEEV, WEEV, and EEEV viral RNA in the brain (1,000-fold to 1,000,000-fold, p < 0.05) at 4 days post-infection (Fig. 4B). We next generated PCDH10-EC1:LDLRAD3-vLA1-VLDLR-LA2-Fc with LALA-PG substitutions in the Fc constant region that abrogate Fc effector functions39. Mice treated with 100 μg of the LALA-PG variant were protected from lethal infection caused by WEEV or VEEV but not EEEV (Fig. 4C). Thus, although the pan-decoy receptors protected against all encephalitic alphaviruses in vivo, the dependence on Fc effector activity varies depending on the virus. This phenotype may be related to the relative neutralizing potency of PCDH10-EC1:LDLRAD3-vLA1-VLDLR-LA2-Fc against the individual viruses (VEEV > WEEV >> EEEV).

Figure 4. PCDH10-EC1:LDLRAD3-vLA1-VLDLR-LA2 protects mice against encephalitic alphavirus challenge.

Figure 4.

Open in a new tab

(A) Survival of mice administered 100 μg of the indicated decoy protein as prophylaxis before challenge with 2 x 102 FFU of EEEV-MADV Argentina 1936, 102 FFU of VEEV ZPC-738, 104 FFU of WEEV McMillan or 104 FFU of WEEV-McMnsp-CBA87 (n = 10, 2 independent experiments). (B) Viral RNA in the brains of challenged mice that were treated with 100 μg of the indicated decoy protein as determined by qRT-PCR in samples harvested at 5 dpi (n = 6, pooled from 2 independent experiments). Data are shown as focus-forming unit (FFU) equivalents derived from a standard curve generated from a viral stock of known titer. The dotted horizontal line indicates the limit of detection. (C) Survival of mice administered 100 μg of the indicated decoy protein before challenge with 2 x 102 FFU of EEEV-MADV Argentina 1936, 102 FFU of VEEV ZPC-738, or 104 FFU of WEEV McMillan or WEEV-McMnsp-CBA87 (n = 8, 2 independent experiments). Statistical analysis in (A and C) was performed using the Log-rank test and in (B) using the Mann-Whitney test; n.s., not significant; *, p < 0.05.

DISCUSSION

In the current study, we leveraged advances in alphavirus receptor identification and structural biology to generate an engineered receptor decoy that broadly neutralizes and protects against lethality in mice after challenge with VEEV, WEEV, or EEEV. We achieved this coverage by designing a stabilized symmetric bispecific scaffold resembling an antibody, which combined structurally distinct molecules (LA and immunoglobulin-like domains) into a single, bivalent decoy Fc-fusion protein. The development of a molecule with broad activity against encephalitic alphaviruses could have utility as prophylaxis or therapy in the absence of laboratory confirmed diagnosis, especially in regions of the world where these viruses co-circulate.

Historically, one strategy to achieve broadly inhibitory antiviral activity has been to screen for antibodies that bind a highly conserved epitope. However, in the absence of readily accessible conserved epitopes on the virion surface of encephalitic alphaviruses, such campaigns have proven challenging. Indeed, to date, a potently neutralizing, pan-reactive antibody against EEEV, WEEV, and VEEV has not yet been identified, although non-neutralizing antibodies have been described that target a highly conserved E1 epitope and protect in vivo through Fc effector function activities40–42. An inhibitory receptor decoy molecule has some theoretical advantages over antibodies. For instance, it may be harder for viruses to escape from decoy molecules and acquire resistance without a substantial fitness cost; escape from the neutralizing decoy likely would compromise the ability of the virus to engage these receptors for efficient entry. Alternatively, de novo protein design tools43–46 could be used to develop smaller, broadly neutralizing proteins or antibody-based molecules that engage conserved epitopes on the viral glycoprotein.

Analysis of the alphavirus E1-E2 cleft region showed substantial variation in electrostatic surface potential with no accessible conserved site, which likely has limited the development of a broadly neutralizing antibody47. A broadly acting receptor decoy, however, is not limited to binding the same sites; indeed, our designed decoy engages distinct sites on each encephalitic alphavirus. Our protein engineering enhanced the breadth of neutralization to EEEV and WEEV and resulted in a decoy molecule that was more inhibitory against VEEV than LDLRAD3-LA1-Fc, due to the addition of a second domain and the introduction of the K58G mutation. The potency of this molecule against VEEV combined with mutations (e.g., YTE [M252Y/S254T/T256E]48 and LS [M428L/N434S])49 to increase half-life in serum could enable the generation of decoy proteins against VEEV with long-acting inhibitory activity. Our proof-of-concept bispecific molecule could serve as a platform for combining structurally distinct receptor antagonists into one molecule that is active against different viruses. Given that LA domain containing molecules have been identified as entry factors for arboviruses spanning multiple genera10–12,16–19, analogous decoy molecules could be designed with broad activity against bunyaviruses or orthoflaviviruses.

Studies with EEEV in mice have demonstrated that Fc effector functions are required for protection with low-affinity but not high-affinity antibodies50,51. Experiments with Mayaro virus, a distantly related arthritogenic alphavirus, showed that even potently neutralizing antibodies require Fc effector functions to mediate protection38. Thus, the differential requirement for harnessing Fc effector functions by the PCDH10-EC1:LDLRAD3-vLA1-VLDLR-LA2-Fc decoy molecule might reflect differences in the mode of pathogenesis of different alphaviruses. Regardless, introduction of mutations in the Fc region that enable enhanced effector functions likely could be another path for augmenting decoy receptor activity in vivo.

We acknowledge several limitations in our study. We have not evaluated the efficacy of our designed molecules against aerosol-mediated infection, which models a bioweapon introduction. We have not assessed the pharmacodynamics of the bispecific platform molecules. Our in vivo assessments were performed in mice. Receptor usage and decoy efficacy could differ in humans. The impact of the introduced mutations into LDLRAD3-LA1-Fc on neutralization potency against different VEEV genotypes was not assessed. Finally, the safety profile of the receptor decoy molecule has not been assessed.

In summary, we utilized structural and functional data to design a receptor-based molecule that neutralizes EEEV, WEEV, and VEEV in cell culture and protects mice from lethal virus challenge. Our broadly neutralizing molecule can be used as a potential therapeutic or prophylaxis against encephalitic viral exposure serving as an alternative to antibody cocktails and vaccines. Our stabilized symmetric bispecific decoy scaffold could be further improved by in silico or saturation mutagenesis-based approaches. Such approaches could enhance their potency against the encephalitic alphaviruses or expand the breadth of inhibition to other viral genera.

Material and Methods

Study design.

The goal of this study was to use our structural knowledge of receptor interactions with different encephalitic alphaviruses to create a single soluble receptor decoy molecule that inhibited infection of EEEV, VEEV, and WEEV in cell culture and in vivo. Structure-guided mutagenesis was iteratively performed to generate decoy molecules that first inhibited interactions with LDLR family members (LDLRAD3 and VLDLR), and then with the WEEV receptor, PCDH10. All data collected was included without exclusion of outliers, and multiple independent experiments were performed to ensure reproducibility. Mice were randomly assigned to treatments, and investigators performing the analyses were not blinded. Sample sizes were chosen based on prior experience in evaluating differences in viral infection in mice.

Cells and viruses.

293T (ATCC, CRL-3216), Vero (ATCC, CCL-81), and BHK21 (ATCC, CCL-10) cells were propagated in high-glucose Dulbecco’s Modified Eagle Media (DMEM, Gibco, Cat# 11965084) supplemented with 10% FBS (Omega Scientific, Cat# FB-01), GlutaMAX (Gibco, Cat# 35050061), 10 mM HEPES pH 7.3 (Gibco, Cat# 15630080), non-essential amino acids (Gibco), and penicillin-streptomycin (Gibco, Cat# 15140122). WEEV McMillan (GenBank: GQ287640), VEEV ZPC738 (GenBank: MF459684.1), VEEV Trinidad Donkey (GenBank: L01442.2), and EEEV-FL93-939 (GenBank: KP282670) were generated from infectious cDNA clones52. EEEV-MADV (Argentina 1936 strain, GenBank: AAC53787.1) was obtained from the World Reference Center for Emerging Viruses and Arboviruses (University of Texas Medical Branch, gift of S. Weaver and K. Plante). SINV chimeric viruses with a GFP reporter were generated by replacing the structural genes in the SINV TR339 infection clone with those of VEEV-TrD, EEEV-FL93, or WEEV strains as described53. A chimeric virus with the non-structural genes of WEEV McMillan and the structural genes of WEEV CBA87 (GenBank: NC_003908.1) was generated by gene synthesis and Gibson assembly and is denoted as WEEV-McM(nsp)-CBA87. All viruses were propagated in BHK21 or BHK21-chicken MXRA87 cells and titered by focus-forming assay on Vero cells.

Mice.

C57BL/6J (Cat #000664, Jackson Laboratories) were maintained in a specific-pathogen-free facility. Four-week-old mice were used for EEEV-MADV infections. Six- to nine-week-old male and female mice were used for all other non-select agent infection (WEEV and VEEV) experiments. Housing and care of laboratory animals was conducted in accordance with guidelines from the NIH Guide for the Care and Use of Laboratory Animals. Animal husbandry and experiments were performed in accordance with guidelines from the Institutional Animal Care and Use Committee at Washington University School of Medicine (Assurance number A3381-01). Virus inoculations were performed under anesthesia induced by ketamine hydrochloride and xylazine.

Select Agent (EEEV FL93-939 and VEEV TrD) experiments with CD-1 mice (Crl:CD1, Charles River) were performed under A-BSL3 conditions using positive-pressure respirators at the University of Pittsburgh Regionl Biocontainment Laboratories under approval of the Institutional Animal Care and Use Committee of the University of Pittsburgh (Assurance number 15066059 and 18073259). Virus inoculations in 6-week-old male and female mice were performed under anesthesia that was induced and maintained either with inhaled isoflurane or ketamine hydrochloride and xylazine (obtained from Division of Comaparative Medicine, Washington Universit School of Medicine). All efforts were made to minimize animal suffering.

Recombinant protein expression and purification.

Fc-fusion constructs and human LRPAP1 (RAP) chaperone gene (residues 1–353) (4:1 ratio) (table S1) were co-transfected with Expifectamine 293 reagent (Thermo Fisher, Cat# A14525) into Expi293 cells (Thermo Fisher, Cat# A14527) according to the manufacturer’s instructions. Bispecific knob-in-hole constructs were transfected with equal parts knob and hole. The symmetric bispecific constructs were transfected at a 1:2 ratio (heavy:kappa chains). Supernatants were harvested four days later, centrifuged, filtered, and Fc-fusion proteins were bound to Protein A Sepharose (Thermo Fisher, Cat# 101042) using a Econo-Pac column (Bio-Rad, Cat# 7321011). The column was washed with 25 volumes of 1x TBS (20 mM Tris pH 8.0, 150 mM NaCl), 50 volumes of high-salt buffer containing ethylenediaminetetraacetic acid (EDTA) (20 mM Tris pH 8.0, 500 mM NaCl, 10 mM EDTA), followed by 25 volumes of 1x TBS + 10 mM CaCl2. Proteins were eluted with Pierce Gentle Ag/Ab Elution Buffer, pH 6.6 (Thermo Fisher, Cat# 21027) and desalted into 1x TBS supplemented with 2 mM CaCl2 using a PD-10 column (Cytiva, Cat# 17085101).

Biolayer interferometry.

Binding experiments were performed on a GatorPlus instrument and analyzed using on-board software (GatorBio). Experiments were performed with 1x TBS supplemented with 1% BSA and 2 mM CaCl2 (Running Buffer). To evaluate binding of VLPs to Fc-fusion proteins, first, anti-human IgG Fc biosensors (GatorBio, Cat# 160003) were incubated with 10 μg/mlof indicated Fc-fusion proteins for 120 sec. The biosensors were then incubated in Running Buffer alone for 60 sec, followed by incubation with indicated VLPs (10 to 20 μg/ml). Data are presented as the signal response during incubation with VLPs without baseline correction. EEEV PE6 and VEEV TC-83 VLPs were gifts of J. Mascola (Vaccine Research Center, NIH, Bethesda, MD), and CHIKV VLPs were a gift from Emergent BioSolutions (Gaithersburg, MD).

Cryo-EM sample preparation and analysis.

VEEV and EEEV VLPs were prepared at a nominal concentration of 0.7 mg/ml in TBS pH 8.0, then incubated for 1 h on ice with 10-fold molar excess of LDLRAD3-vLA1-VLDLR-LA2, which was generated by cleavage of Fc-fusion proteins with GlySERIAS (Genovis, Cat# A0-GS6-020) followed by capture of released Fc fragments with Protein A Sepharose (Thermo Fisher, Cat# 101042) and concentration of the cleaved fragments with an Amicon 3 kDa centrifugal filter (Millipore-Sigma, Cat# UFC5003). Solutions were applied to glow-discharged lacey carbon grids (Ted Pella #01895-F) then flash-frozen in liquid ethane using a Vitrobot Mark IV (Thermo Fisher).

Grids were loaded into a Cs-corrected FEI Titan Krios 300 kV microscope equipped with a Falcon 4 direct electron detector, and then imaged at a nominal magnification of 59,000 × resulting in a calibrated pixel size of 1.081 Å. Movies were collected in electron event representation (EER) format with a total dose of 47-48 e−/Å2/movie (approximately 4.5 e−/Å2/s over 8-9 sec acquisition). EER movies were binned into fractions (approximately 1.0 e−/Å2/f) and then pre-processed via patch motion and patch CTF correction in cryoSPARC v4.4. Particles were selected using a template picker and then cleaned via two- and three-dimensional classification. VLPs were reconstructed via homogeneous non-uniform refinement with icosahedral symmetry imposed. Following symmetry expansion, extracted asymmetric units were subjected to several rounds of focused refinement in cryoSPARC v4.4 (table S2)54. We rigid-body docked previously generated atomic models of EEEV in complex with VLDLR LA(1–2) (PDB: 8UFC29) and VEEV in complex with LDLRAD3 LA1 (PDB: 7FFF31) into the corresponding maps. No additional refinement was performed.

Neutralization studies.

To assess the neutralizing activity of Fc-fusion proteins, HEK293T cells were plated (2.5 x 104 per well) in a 96-well plate and allowed to adhere overnight. The following day, viruses (multiplicity of infection [MOI] of 0.5) were incubated with Fc-fusion proteins for 1 h at 37°C in DMEM supplemented with 2% FBS before inoculation. Cells were harvested 16 h later and subjected to analysis on an iQue3 flow cytometer (Sartorius) and analyzed with on-board iQue Forecyt Software (Sartorius).

Mouse experiments.

Mice were administered a 100-μg dose of Fc-fusion proteins via intraperitoneal injection 24 h prior to subcutaneous footpad inoculation with 104 FFU of WEEV McMillan, 104 FFU of WEEV McMillan(nsp)-CBA87, 102 FFU of VEEV ZPC-738, 103 FFU of VEEV TrD or FL93-939, or 2 x 102 MADV. Mice were monitored daily for survival. At 5 dpi, some cohorts were euthanized, and after extensive perfusion with PBS (Gibco, Cat# 14190250), tissues were collected and stored at −80°C prior to processing. Tissues were homogenized with a MagNA Lyser (Roche), and viral RNA was extracted with a MagMAX-96 Viral RNA extraction kit (Thermo Fisher, Cat# AMB18365) using a Kingfisher Flex instrument (Thermo Fisher) according to the manufacturer’s instructions. Amounts of viral RNA were determined using a TaqMan RNA-to-Ct 1-Step Kit (Thermo Fisher, Cat# 4392938) on a QuantStudio 6 (Thermo Fisher) using the following primer and probe sets: WEEV-forward 5′-AGATATTGCCCAATCCAGAAA-3′, WEEV-reverse 5′ -TATGCGCCTCTGAAGGAAATAG-3′, WEEV-probe 5′-/56-FAM/AAGCAATTA/ZEN/CAGCGGAGCGACTCA/3IABKFQ/, MADV nsP3 forward: 5′-ATGAAGCGGGTGCGTATATC-3′, EEEV-MADV nsP3 reverse: 5′-CAGAATAGGGTGCTGGAGTTT-3′, EEEV-MADV probe 5′-/56-FAM/ATTTCTGTT/ZEN/GCAGATGCCCTTGCC/3IABkFQ/−3′, VEEV ZPC738 nsP3 forward 5’ -CAAGTCGAGGCAGACATTCA-3’, VEEV ZPC738 nsP3 reverse 5’ -CAGGGTGTCAAGGATGGATAAA-3’, and VEEV ZPC738-probe 5’-/56-FAM/TGGTCCATT/ZEN/ CCTCATGCATCCGAC/3IABkFQ/−3’. FFU equivalents were determined by processing of a viral stock of known titer for amounts of viral RNA.

Statistical analysis.

Individual-level data are presented in data file S1. Statistical significance was determined using Prism Version 10 (GraphPad). Statistical analysis of viral infection was evaluated by one-way ANOVA with Dunnett’s post-test or Mann-Whitney test. Analysis or mortality was determined by the log-rank test. The number of independent experiments are indicated in the figure legends. Technical replicates are plotted to show variance, but statistical tests were conducted on biological replicates.

Supplementary Material

aec2923 SupplementalMaterial Final
Data File S1

Acknowledgements.

We thank B. Readnour, B. Summers, and K. Basore at the Washington University Center for Cellular Imaging for cryo-EM sample preparation and data acquisition, as well as D. Fremont and members of his laboratory for experimental discussions.

Funding.

The studies were supported by NIH grants R01 AI141436 (to M.S.D.), R01 AI153209 (to M.S.D.), U19 AI142790 (to M.S.D.), and K08 AI187716 (to S.R.). This effort was sponsored by the U.S. Government (Defense Threat Reduction Agency) under contracts W15QKN-16-9-1002 and MCDC2103-01 (to W.B.K.) between the Medical CBRN Defense Consortium and the Government.

Competing Interests.

M.S.D. is a consultant for Inbios, Moderna, IntegerBio, Merck, GlaxoSmithKline, and Akagera Medicines. The Diamond laboratory has received unrelated funding from Moderna (to M.S.D.). W.B.K. is a co-founder of Advanced Virology. All other authors declare no competing interests.

Data, Code and Materials availability.

All data associated with this study are reported in the manuscript, the Electron Microscopy Data Bank (EMDB) public repository with indicated accession numbers, or in the supplementary data file. The electron density maps are deposited in the EMDB with the following accession codes: EMD-74157 (VEEV in complex with LDLRAD3-vLA1-VLDLR-LA2) and EMD-74158 (EEEV in complex with LDLRAD3-vLA1-VLDLR-LA2). Plasmids, proteins, sequences, viruses, and cell lines will be shared by the contacting author (mdiamond@wustl.edu) upon completion of a materials transfer agreement or will be supplied upon reasonable request. All materials (e.g., viruses, antibodies, cell lines, and proteins) used or generated in this study are commercially available or will be supplied upon reasonable request.

References and Notes

  • 1.Meyer KF, Haring CM & Howitt B The etiology of Epizootic encephalomyelitis of horses in the San Joaquin valley, 1930. Science 74, 227–228 (1931). [DOI] [PubMed] [Google Scholar]
  • 2.Merrill MH, Lacaillade CW Jr, & Broeck CT Mosquito transmission of equine encephalomyelitis. Science 80, 251–252 (1934). [DOI] [PubMed] [Google Scholar]
  • 3.Beck CE & Wyckoff RW Venezuelan equine encephalomyelitis. Science 88, 530 (1938). [DOI] [PubMed] [Google Scholar]
  • 4.Bell JR, Kinney RM, Trent DW, Strauss EG & Strauss JH An evolutionary tree relating eight alphaviruses, based on amino-terminal sequences of their glycoproteins. Proc. Natl. Acad. Sci. U. S. A. 81, 4702–4706 (1984). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Kafai NM et al. Neutralizing antibodies protect mice against Venezuelan equine encephalitis virus aerosol challenge. J. Exp. Med. 219, (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Ma H et al. LDLRAD3 is a receptor for Venezuelan equine encephalitis virus. Nature 588, 308–314 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Palakurty S et al. The VLDLR entry receptor is required for the pathogenesis of multiple encephalitic alphaviruses. Cell Rep. 43, 114809 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Li W et al. Shifts in receptors during submergence of an encephalitic arbovirus. Nature 1–8 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Clark LE et al. VLDLR and ApoER2 are receptors for multiple alphaviruses. Nature 602, 475–480 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ganaie SS et al. Lrp1 is a host entry factor for Rift Valley fever virus. Cell 184, 5163–5178.e24 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Schwarz MM et al. Oropouche orthobunyavirus infection is mediated by the cellular host factor Lrp1. Proc. Natl. Acad. Sci. U. S. A. 119, e2204706119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Huerta V et al. The low-density lipoprotein receptor-related protein-1 is essential for dengue virus infection. Viruses 16, 1692 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Chong Z et al. Multiple LDLR family members act as entry receptors for yellow fever virus. Nature 1–10 (2025). [DOI] [PubMed] [Google Scholar]
  • 14.Herdy B, Snyers L, Reithmayer M, Hinterdorfer P & Blaas D Identification of the human rhinovirus serotype 1A binding site on the murine low-density lipoprotein receptor by using human-mouse receptor chimeras. J. Virol. 78, 6766–6774 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Finkelshtein D, Werman A, Novick D, Barak S & Rubinstein M LDL receptor and its family members serve as the cellular receptors for vesicular stomatitis virus. Proc. Natl. Acad. Sci. U. S. A. 110, 7306–7311 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Ritter M et al. The low-density lipoprotein receptor and apolipoprotein E associated with CCHFV particles mediate CCHFV entry into cells. Nat. Commun. 15, 4542 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Xu Z-S et al. LDLR is an entry receptor for Crimean-Congo hemorrhagic fever virus. Cell Res. 34, 140–150 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Monteil VM et al. Crimean-Congo haemorrhagic fever virus uses LDLR to bind and enter host cells. Nat Microbiol (2024) doi: 10.1038/s41564-024-01672-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Bhaskar M, Satheesan A & Basu A Low-density Lipoprotein Receptor is an important host factor in flaviviral entry and replication in neurons. Biochem. Biophys. Res. Commun. 743, 151160 (2025). [DOI] [PubMed] [Google Scholar]
  • 20.Zhai X et al. LDLR is used as a cell entry receptor by multiple alphaviruses. Nat. Commun. 15, 622 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Mittler E et al. LRP8 is a receptor for tick-borne encephalitis virus. Nature (2025) doi: 10.1038/s41586-025-09500-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Li P et al. LRP8 is an entry receptor for tick-borne encephalitis viruses. Proc. Natl. Acad. Sci. U. S. A. 122, e2525771122 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Andersen OM et al. Identification of the minimal functional unit in the low density lipoprotein receptor-related protein for binding the receptor-associated protein (RAP). A conserved acidic residue in the complement-type repeats is important for recognition of RAP. J. Biol. Chem. 275, 21017–21024 (2000). [DOI] [PubMed] [Google Scholar]
  • 24.Fass D, Blacklow S, Kim PS & Berger JM Molecular basis of familial hypercholesterolaemia from structure of LDL receptor module. Nature 388, 691–693 (1997). [DOI] [PubMed] [Google Scholar]
  • 25.Raju S et al. Structural basis for plasticity in receptor engagement by an encephalitic alphavirus. Cell (2025) doi: 10.1016/j.cell.2025.02.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Ma B et al. Structural basis for the recognition of two different types of receptors by Western equine encephalitis virus. Cell Rep. 44, 115724 (2025). [DOI] [PubMed] [Google Scholar]
  • 27.Cao D, Ma B, Cao Z, Zhang X & Xiang Y Structure of Semliki Forest virus in complex with its receptor VLDLR. Cell 186, 2208–2218.e15 (2023). [DOI] [PubMed] [Google Scholar]
  • 28.Cao D et al. The receptor VLDLR binds Eastern Equine Encephalitis virus through multiple distinct modes. Nat. Commun. 15, 6866 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Adams LJ et al. Structural and functional basis of VLDLR usage by Eastern equine encephalitis virus. Cell 0, (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Basore K et al. Structure of Venezuelan equine encephalitis virus in complex with the LDLRAD3 receptor. Nature 598, 672–676 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Ma B, Huang C, Ma J, Xiang Y & Zhang X Structure of Venezuelan equine encephalitis virus with its receptor LDLRAD3. Nature 598, 677–681 (2021). [DOI] [PubMed] [Google Scholar]
  • 32.Yang P et al. Structural basis for VLDLR recognition by eastern equine encephalitis virus. Nat. Commun. 15, 6548 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Fan X et al. Molecular basis for shifted receptor recognition by an encephalitic arbovirus. Cell 188, 2957–2973.e28 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Yang Y et al. PCDH10 is a neuronal receptor for western equine encephalitis virus. Cell Res. 34, 802–805 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ridgway JB, Presta LG & Carter P “Knobs-into-holes” engineering of antibody CH3 domains for heavy chain heterodimerization. Protein Eng. 9, 617–621 (1996). [DOI] [PubMed] [Google Scholar]
  • 36.Janeway CA Jr, Travers P, Walport M & Shlomchik MJ The structure of a typical antibody molecule. in Immunobiology: The Immune System in Health and Disease. 5th edition (Garland Science, 2001). [Google Scholar]
  • 37.Mazor Y et al. Improving target cell specificity using a novel monovalent bispecific IgG design. MAbs 7, 377–389 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Chen W et al. Improving the CH1-CK heterodimerization and pharmacokinetics of 4Dm2m, a novel potent CD4-antibody fusion protein against HIV-1. MAbs 8, 761–774 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Schlothauer T et al. Novel human IgG1 and IgG4 Fc-engineered antibodies with completely abolished immune effector functions. Protein Eng. Des. Sel. 29, 457–466 (2016). [DOI] [PubMed] [Google Scholar]
  • 40.Williamson LE et al. Therapeutic alphavirus cross-reactive E1 human antibodies inhibit viral egress. Cell 184, 4430–4446.e22 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Sutton MS et al. Vaccine elicitation and structural basis for antibody protection against alphaviruses. Cell 186, 2672–2689.e25 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Kim AS et al. Pan-protective anti-alphavirus human antibodies target a conserved E1 protein epitope. Cell 184, 4414–4429.e19 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Cao L et al. Design of protein-binding proteins from the target structure alone. Nature 605, 551–560 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Lv X et al. De novo design of mini-protein binders broadly neutralizing Clostridioides difficile toxin B variants. Nat Commun 15, 1–14 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Watson JL et al. De novo design of protein structure and function with RFdiffusion. Nature 620, 1089–1100 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Abramson J et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630, 493–500 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Wang N et al. LDL receptor in alphavirus entry: structural analysis and implications for antiviral therapy. Nat. Commun. 15, 4906 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Dall’Acqua WF, Kiener PA & Wu H Properties of human IgG1s engineered for enhanced binding to the neonatal Fc receptor (FcRn). J. Biol. Chem. 281, 23514–23524 (2006). [DOI] [PubMed] [Google Scholar]
  • 49.Zalevsky J et al. Enhanced antibody half-life improves in vivo activity. Nat. Biotechnol. 28, 157–159 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Callahan V et al. High binding potency overcomes the requirement of Fc effector functions for broadly reactive anti-alphavirus antibodies. Sci. Transl. Med. 17, eadt9853 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Earnest JT et al. Neutralizing antibodies against Mayaro virus require Fc effector functions for protective activity. J. Exp. Med. 216, 2282–2301 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Gardner CL et al. In vitro and in vivo phenotypes of Venezuelan, eastern and western equine encephalitis viruses derived from cDNA clones of human isolates. Viruses 15, 5 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Sun C, Gardner CL, Watson AM, Ryman KD & Klimstra WB Stable, high-level expression of reporter proteins from improved alphavirus expression vectors to track replication and dissemination during encephalitic and arthritogenic disease. J. Virol. 88, 2035–2046 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Punjani A, Rubinstein JL, Fleet DJ & Brubaker MA cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017). [DOI] [PubMed] [Google Scholar]
  • 55.Saphire EO et al. Crystal structure of a neutralizing human IGG against HIV-1: a template for vaccine design. Science 293, 1155–1159 (2001). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

aec2923 SupplementalMaterial Final
NIHMS2181888-supplement-aec2923_SupplementalMaterial_Final.pdf (7.8MB, pdf)
Data File S1
NIHMS2181888-supplement-Data_File_S1.xlsx (615KB, xlsx)

Data Availability Statement

All data associated with this study are reported in the manuscript, the Electron Microscopy Data Bank (EMDB) public repository with indicated accession numbers, or in the supplementary data file. The electron density maps are deposited in the EMDB with the following accession codes: EMD-74157 (VEEV in complex with LDLRAD3-vLA1-VLDLR-LA2) and EMD-74158 (EEEV in complex with LDLRAD3-vLA1-VLDLR-LA2). Plasmids, proteins, sequences, viruses, and cell lines will be shared by the contacting author (mdiamond@wustl.edu) upon completion of a materials transfer agreement or will be supplied upon reasonable request. All materials (e.g., viruses, antibodies, cell lines, and proteins) used or generated in this study are commercially available or will be supplied upon reasonable request.

RESOURCES