(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Human herpesvirus 8 molecular mimicry of ephrin ligands facilitates cell entry and triggers EphA2 signaling

['Taylor P. Light', 'Department Of Materials Science', 'Engineering', 'Johns Hopkins University', 'Baltimore', 'Maryland', 'United States Of America', 'Institute For Nanobiotechnology', 'Delphine Brun', 'Department Of Virology']

Date: 2021-10

Human herpesvirus 8 (HHV-8) is an oncogenic virus that enters cells by fusion of the viral and endosomal cellular membranes in a process mediated by viral surface glycoproteins. One of the cellular receptors hijacked by HHV-8 to gain access to cells is the EphA2 tyrosine kinase receptor, and the mechanistic basis of EphA2-mediated viral entry remains unclear. Using X-ray structure analysis, targeted mutagenesis, and binding studies, we here show that the HHV-8 envelope glycoprotein complex H and L (gH/gL) binds with subnanomolar affinity to EphA2 via molecular mimicry of the receptor’s cellular ligands, ephrins (Eph family receptor interacting proteins), revealing a pivotal role for the conserved gH residue E52 and the amino-terminal peptide of gL. Using FSI-FRET and cell contraction assays, we further demonstrate that the gH/gL complex also functionally mimics ephrin ligand by inducing EphA2 receptor association via its dimerization interface, thus triggering receptor signaling for cytoskeleton remodeling. These results now provide novel insight into the entry mechanism of HHV-8, opening avenues for the search of therapeutic agents that could interfere with HHV-8–related diseases.

The structures of several Eph receptor–ephrin ligand complexes have been determined [ 27 , 28 , 33 ], but how viral antigens such as HHV-8 gH/gL interact with EphA2 was completely unknown until recently. The 3.2-Å X-ray structure of a HHV8 gH/gL-EphA2 complex was published while we were preparing this manuscript [ 34 ]. Our goal has been to explore the events that emulate the early stages of HHV-8 entry. We sought to obtain the structural details on the gH/gL-EphA2 complex and to determine if and how HHV-8 gH/gL affects assembly of EphA2 receptors in the membranes of living cells, taking advantage of the Fully Quantified Spectral Imaging–Förster Resonance Energy Transfer (FSI-FRET) system that allows quantification of lateral interactions of membrane proteins in vivo [ 35 ]. We report here a 2.7-Å resolution X-ray structure of the HHV-8 gH/gL ectodomain bound to the LBD of EphA2 together with results of structure-guided mutagenesis and cell-based studies. Based on our analyses and the similarities we observed between the binding modes of gH/gL and the ephrin ligand to EphA2, we provide evidence that this structural similarity extends into functional mimicry. The results presented here now lay a path for further exploration of downstream events and the investigation of whether HHV-8 activation of Eph receptors may play a role beyond ensuring a productive infection, for example, contributing to virus oncogenicity/oncogenic transformation of the cell.

At the molecular level, as in the case of other receptor tyrosine kinases, ephrin ligand binding induces oligomerization of Eph receptors, promoting trans-phosphorylation and signal transduction into the cell (reviewed in [ 25 ]). Structural and functional studies revealed that ephrin ligand binding to Eph receptors results first in formation of tetramers made of 2 “Eph-ephrin” complexes within which the Eph receptors form dimers stabilized via a specific interface in their LBD called the dimerization interface (DIN) (reviewed in [ 26 ]; S2 Fig ). As ligand concentration increases, such Eph-ephrin tetramers polymerize into larger clusters via a surface in the downstream CRD, which is referred to as the clustering surface (CIN) [ 27 – 29 ]. The cellular response to EphA2 receptor activation is ligand and cell type dependent and modulated by factors such as size and type of the EphA2 receptor oligomers, the spatial distribution of the receptor in the membrane [ 30 ], residues in the intracellular domain that are phosphorylated, to just name some [ 16 ]. Different ligands (monomeric, dimeric ephrin-A1, and agonist or antagonist peptides) were shown to stabilize distinct dimeric or oligomeric EphA2 receptor assemblies ( S2 Fig ), further indicating that the signaling properties may be defined by the nature of the EphA2 dimers and oligomers [ 31 , 32 ].

All Eph receptors contain an elongated ectodomain made of—as beads on a string—a ligand-binding domain (LBD), a cysteine-rich domain (CRD), and 2 fibronectin (FN) domains, followed by a transmembrane anchor, a short juxtamembrane region containing several conserved tyrosine residues, an intracellular Tyr kinase domain, a sterile alpha motif (SAM) that has a propensity to oligomerize, and a PDZ domain involved in protein–protein interactions [ 22 ] ( S2 Fig ). We employ the accepted nomenclature for the secondary structure elements for the LBD of Eph receptors [ 23 ] and ephrin ligands throughout the text [ 24 ] ( S3 Fig ). In both cases, single letters designate β-strands and helices, and double letters are used to label loops that connect the secondary structure elements. To avoid confusion, we use superscripts to indicate the molecule the residue or feature is ascribed to (R103 EphA2 , E119 ephrin , GH EphA2 , GH ephrin , etc.).

The physiological ligands of Eph receptors are membrane-tethered proteins called ephrins (acronym for Eph family receptor interacting proteins) ( S2 Fig ). Eph receptor–ephrin ligand interactions mediate short-distance cell–cell communications and lead to cytoskeleton rearrangements and rapid changes in cell mobility and/or morphology [ 16 ]. Some of the typical outcomes of ephrin-A1 ligand activation of EphA2 receptor are cell retraction [ 17 – 19 ] and endocytosis of receptor–ligand complexes [ 20 ]. These processes are especially active and important during development, and in adulthood, many of the same circuits get repurposed for functions in bone homeostasis, angiogenesis, and synaptic plasticity (reviewed in [ 16 ]). Since motility and angiogenesis contribute to tumorigenesis and other pathologies, Eph receptors and ephrin ligands are in the spotlight as targets for therapeutic intervention [ 21 ].

EphA2, where Eph stands for erythropoietin-producing human hepatocellular carcinoma cell line, was identified as HHV-8 entry receptor by Hahn and colleagues who showed that deletion of the EphA2 gene abolished infection of endothelial cells and that binding of gH/gL to EphA2 on cells led to increased EphA2 phosphorylation and endocytosis facilitating viral entry [ 10 ]. The presence of the intracellular kinase domain was found to be important for HHV-8 entry in epithelial 293 cells [ 10 ]. In this respect, HHV-8 gH/gL does not play the role of a classical herpesvirus receptor binding protein that would directly activate gB upon binding to a cellular receptor to induce fusion of the viral and plasma membranes, such as gD in alphaherpesviruses or gp42 in Epstein–Barr virus (EBV), for example (reviewed in [ 8 ]). HHV-8 gH/gL instead activates EphA2 receptors that initiate signaling pathways leading to rapid internalization of the virus and cytoskeletal rearrangements that create a cellular environment conducive for the virus and capsid intracellular transport [ 11 ]. HHV-8 binds with the highest affinity to EphA2 and less to the related EphA4 and EphA7 receptors [ 10 , 12 , 13 ]. In addition, the EphA2 receptor serves as a receptor for 2 other gammaherpesviruses—human herpesvirus 4 (HHV-4, also known as EBV) and rhesus monkey rhadinovirus [ 14 , 15 ]. The interactions are in all cases established via gH/gL.

Behind the ability of HHV-8 to spread to diverse tissues lies its wide tropism demonstrated in vivo for epithelial and endothelial cells, fibroblasts, B and T lymphocytes, monocytes, macrophages, and dendritic cells (reviewed in [ 5 ]). The major route of HHV-8 entry is via endocytosis [ 6 ] ( S1 Fig ). As other herpesviruses, HHV-8 first attaches to cells via its glycoproteins that protrude from the virus surface and engage in numerous low-affinity interactions with ubiquitous cellular factors such as heparan sulfate proteoglycans [ 7 ]. The virus is trafficked toward the endosomal compartment, where the capsids are released into cytosol upon merger of the viral and endosomal membranes [ 6 ] ( S1 Fig ). The membrane fusion process is mediated by the envelope glycoprotein B (gB) and the noncovalent heterodimer made of glycoproteins H and L (gH/gL), which constitute the conserved core fusion machinery of all herpesviruses. The gB is the fusogen protein, while gH/gL plays a role in the regulation of gB activity [ 8 ]. The current model posits that upon a fusion trigger, gH/gL in a still unknown way relays the signal and switches gB fusion activity on, setting membrane fusion in motion [ 9 ]. What is particular to HHV-8 is the simultaneous employment of several viral glycoproteins—HHV-8–specific K8.1A glycoprotein, as well as the core fusion machinery components—that engage diverse cellular receptors (gB binds to integrins and DC-SIGN, gH/gL to EphA receptors), increasing the HHV-8 target repertoire and providing the virus with a set of tools for well-orchestrated entry (reviewed in [ 6 ]).

Human herpesvirus 8 (HHV-8), also known as Kaposi sarcoma (KS)-associated virus, is a member of Rhadinovirus genus that belongs to the Gammaherpesvirinae subfamily of Herpesviridae [ 1 ]. HHV-8 is an oncogenic virus and etiological agent of KS, malignancy of endothelial cells named after the Hungarian dermatologist who first described the disease in 1872 [ 2 ]. Because KS has different clinical manifestations, 2 main forms are distinguished—the classic KS that is a relatively indolent and rare tumor, appearing as skin lesions mostly in elderly men, and the epidemic or HIV-associated KS, an aggressive form that spreads extensively through skin, lymph nodes, intestines, and lungs. The KS affects up to 30% of untreated HIV–positive individuals [ 3 ] and is, nowadays, one of the most frequent malignancies in men and children in subequatorial African countries [ 4 ].

Results

The gH/gL-EphA2 LBD complex structure Recombinant HHV-8 gH/gL ectodomains and EphA2 LBD (Fig 1A) were expressed in insect cells, and the proteins were purified as described in detail in Materials and methods. To maximize the tertiary complex (gH/gL bound to EphA2 LBD) formation, the gH/gL was mixed with an excess of LBD, which was then removed by size exclusion chromatography (SEC). Multiangle light scattering (MALS) measurements demonstrated a 1:1:1 tertiary complex stoichiometry for gL/gH bound to EphA2 LBD, as well as to the EphA2 ectodomain (S4 Fig). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Schematic representation of HHV-8 gH, gL, and EphA2 and structure of the gH/gL-EphA2 LBD complex. (A) Schematic representation of the EphA2 receptor, HHV-8 gL, and gH, highlighting the protein segments that were expressed as recombinant proteins for crystallization and the residues resolved in the structure. The short fragments that could not be built in EphA2 LBD and gH because of the poor electron density are marked with dotted lines and labeled as breaks (b) (the missing residues are listed in S1 Text). The disulfide bonds are indicated with yellow numbers, and N-linked glycosylation sites with hexagons (green with black border—built in our structure; green with gray border—built in the PBD accession code 7CZF [34]; white, with black borders—the remaining predicted sites). Signal peptides at the start of each protein are represented as white boxes with gray lines, transmembrane anchor domains in EphA2 and gH as dark gray boxes, and double strep tag for affinity purification on gL and EphA2 LBD as half circles. (B) The structure of the tertiary complex is represented as molecular surface and cartoon model (EphA2 LBD in purple, gL in blue, and gH in gray). The N and C termini of each protein are labeled with letters “N” and “C,” respectively. The 4 domains of gH are marked with roman numbers on the left side, and putative locations of the viral and cellular membranes with dashed arrows (black and purple, respectively). The hinge/linker region on gH is indicated with a gray arrow, and putative position of the unresolved J helix in the LBD with a cyan * symbol. Disulfide bonds are represented with yellow sticks. CRD, cysteine-rich domain; HHV-8, human herpesvirus 8; gH/gL, glycoproteins H and L; LBD, ligand-binding domain; SAM, sterile alpha motif. https://doi.org/10.1371/journal.pbio.3001392.g001 The tertiary complex forms an extended structure 15 nm long and around 4.6 nm across its widest part, in the gH region (Fig 1B). EphA2 LBD adopts a jelly roll fold as originally described [33]—its N- and C-termini point in the same direction and away from the gH/gL binding site, consistent with the expected location of the remaining EphA2 domains. Two antiparallel 5-stranded β-sheets pack into a compact β-sandwich, with loops of different lengths connecting the strands. The HIEphA2 loop is well ordered and forms the DIN, while the JKEphA2 loop, which carries a short J′ helix, is not resolved in our tertiary complex structure, likely due to its already reported structural plasticity [36] and/or displacement by gL (Fig 1B). Apart from the JKEphA2 loop, the EphA2 LBD does not change conformation upon binding to gH/gL. Clear electron density was observed at 4 N-linked glycosylation sites (N46gH, N267gH, N688gH, and N118gL) allowing placement of 1 or 2 N-acetylglucosamine residues. The gH/gL complex has an architecture already described for other herpesvirus orthologs—the γ-herpesvirus EBV gH/gL [37], β-herpesvirus human CMV [38], and α-herpesviruses HSV-2, PrV, and VZV [39–41]. The N-terminal domain I (DI) of gH is separated by a linker or hinge helix from the rest of the ectodomain, i.e., domains II, III, and the membrane-proximal domain IV (Fig 1B). The HHV-8 gH/gL resembles the most its EBV counterpart, consistent with the highest sequence conservation between the two, followed by the β-herpesvirus CMV gH/gL complex and less so the α-herpesvirus complexes (S5 Fig). The RMSD values and Z-scores calculated from the superimposition of individual gH domains and gL are given in S5B Fig (superimposing the entire gH/gL ectodomains is not informative because of the different orientations of the domains with respect to each other). Su and colleagues reported the crystal structure of the same tertiary complex (PDB accession code 7CZF) [34] while we were preparing our manuscript. The 2 structures are very similar, with the RMSD value of 6.4 Å for the superimposition of the 2 tertiary complexes. The relatively high RMSD value stems from the disposition of gH in respect to the EphA2 LBD and gL end of the molecule that align very well (RMSD <1 Å) (S5B Fig). These movements are likely a consequence of the flexible hinge helix connecting domains I and II of gH (Fig 1B) and could also be influenced by the different packing of molecules within the 2 crystal lattices (P2 1 2 1 2 1 and C222 1 for the PDB: 7CZF [34] and our structure PDB: 7B7N, respectively).

Binding interface between gH/gL and EphA2 The EphA2 LBD and gH/gL form an intricate interface structure made of a 7-stranded mixed β-sheet containing strands contributed by all 3 proteins. The N-terminal segment of gH co-folds with gL forming a mixed 5-stranded β-sheet composed of 2 gH and 3 gL β-strands. The third gL β-strand further engages in contacts with the β-strand D of EphA2 (Fig 2A). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. The binding interface between EphA2 LBD and HHV-8 gH/gL. (A) The mixed β-sheet formed by β-strands of gH, gL, and EphA2. The strands in gH and gL are labeled as β number , while the EphA2 LBD strands are marked using the single-letter nomenclature assigned for the first solved structure of the EphB2 LBD 1KGY [23]. The same coloring scheme as in Fig 1B is applied. The inlet illustrates the organization of the interacting structural elements—the channel formed by the EphA2 and gL strands (“roof”) that accommodates the gL N-terminal “tail,” reinforced by polar interactions between R103EphA2 and E52gH (“base”) (right panel). (B) Locations of the point mutations introduced in EphA2 (R103A) and gH (E52A, E52R), and N-linked glycosylation sites in EphA2 (N57, A190N) and gL (Q30N, D68N) are indicated, and their side chains are shown as sticks. The same coloring scheme as in Fig 1B is applied. (C) Sensorgrams recorded for WT EphA2 ectodomain of LBD binding to immobilized gH/gL by BLI. A series of measurements using a range of concentrations for EphA2 ectodomain and LBD, respectively, was carried out to obtain the Kd for the WT proteins. Experimental curves (colored traces) were fit using a 1:1 binding model (black traces) to derive equilibrium K d values. (D) Sensorgrams recorded for EphA2 variants binding to immobilized gH/gL variants by BLI. Single experimental curves obtained for EphA2 ectodomain concentration of 62.5 nM plotted to show the effect of the EphA2 mutations, gL mutations, and gH mutations on binding, respectively. The underlying data for panels (C) and (D) can be found in S1 Data. BLI, Biolayer interferometry; gH/gL, glycoproteins H and L; HHV-8, human herpesvirus 8; LBD, ligand-binding domain; WT, wild-type. https://doi.org/10.1371/journal.pbio.3001392.g002 gL binds to the EphA2 LBD via its N-terminal segment (residues 21 to 30) and residues from its β2 and β3 strands (Fig 2A). The full list of contact residues is given in S2 Table and is represented in S6 Fig. The gL N-terminal segment is restrained by C26 and C27 that form disulfide bonds with C74 and C54, respectively. Immediately upstream this anchoring point there is an elongated, hydrophobic “tail” (residues 21 to 25) that inserts into a hydrophobic channel formed by the EphA2 strands D and E and gL strands β2 and β3 (the “roof”). The buried surface area for the gL residues 19 to 32 is around 480 Å2. Of the 14 hydrogen bonds formed between gL and EphA2, 7 are contributed by the gL N-terminal segment, 6 by strand β3, and 1 by the C-terminal η4 N128. Below the gL “tail,” the single gH residue that makes contacts with EphA2—E52gH—forms a salt bridge with R103EphA2, clamping the bottom of the tunnel (the “base”). E52gH is also involved in polar interactions with residues V22gL, H47gL, and F48gL, thus being a center point (a hub) interlaying gL and EphA2 (S2 Table).

Biolayer interferometry (BLI) analyses of EphA2 and gH/gL interactions in solution To investigate the role of the gH/gL and EphA2 residues implicated in the interactions observed in the crystal structure, we tested a series of mutants that were conceived to induce large perturbations in gL or EphA2 by introducing N-glycosylation sites. We resorted to such drastic changes because most EphA2 point mutations already tested in immunoprecipitation assays had only moderate effects on gH/gL binding [42]. The point mutation R103AEphA2 has been reported to abolish the binding to gH/gL [34] and served as a positive control. Since E52gH is the only gH residue contacting EphA2, we also introduced point mutations E52AgH and E52RgH to specifically target this site. The variants with the following N-glycosylation sites were generated by substitution of residues to introduce the N-glycosylation NXS/T motifs at N57EphA2 (M59SEphA2), N190EphA2 (A190NEphA2, L192S EphA2), N30gL (Q30N gL), or N68 gL (D68N gL). Variants containing point mutations were R103AEphA2, E52AgH bound to gL (E52AgH/ gL) and E52RgH/ gL (Fig 2B). The recombinant proteins were expressed in mammalian cells. The introduced sites N30gL and N68 gL were glycosylated as clearly observed by gL shift to a higher molecular weight on SDS-PAGE gels (S7 Fig, S1 Raw Data), while the change in the migration was harder to detect for EphA2 ectodomains possibly because its larger size and small difference introduced by an additional glycosylation. The N190EphA2 mutation was already reported to perturb the interactions with ephrin ligands due to the additional glycosylation site [27]. All gH/gL constructs were engineered so that gL contained a strep tag for complex purification, as before, and gH contained a histidine tag at C terminus for immobilization onto BLI sensors via the end distal to the EphA2 binding site (S8 Fig). Further details on protein production and BLI parameters are given in Materials and methods. We determined the dissociation constant (Kd) <1 nM by doing a series of BLI measurements for the wild-type (WT) gH/gL binding to the EphA2 ectodomain (res. 27 to 534) or EphA2 LBD (res. 27 to 202) (Fig 2C). The low Kd observed for the WT proteins was dominated by a slow k off rate. We obtained a Kd in the subnanomolar range when the measurements were done at pH 5.5 (S9A Fig), or when the system was inverted, i.e., EphA2 LBD or ectodomains were immobilized via a histidine-tag to the sensor, and gH/gL was in solution (S9B Fig). Each of the 3 mutations introduced in EphA2 significantly reduced the binding as anticipated (Fig 2D). The Q30NgL mutation in the gL N-terminal segment also diminished binding, consistent with the presence of a carbohydrate at this position blocking the interactions with the strand DEphA2 and DEEphA2 loop. Introduction of the N-linked carbohydrate at residue N68gL in its β2-β3 turn did not affect binding as expected, because of its location in an exposed loop proximal to the binding site (Fig 2B). The E52RgH/gL and E52AgH/gL variants resulted in weaker or absence of interactions with EphA2 ectodomains, respectively (Fig 2D). These results demonstrated that the binding interface between EphA2 and gH/gL seen in the crystal is in agreement with the one mapped by measurements in solution.

The gH/gL molecular mimicry of ephrin-A ligands The EphA2 binding site for ephrin-A1 ligand and gH/gL largely overlap, with the former including a more extensive surface area and a larger number of contacts established by EphA2 β-strands D and E, CD and DE loops, as well as the LM loop (Figs 3 and S3). We noticed that the gH β-turn (SIELEFNGT) that includes E52gH and F53gH residues (underlined) carries resemblance to the motif located within a GHephrin-A1 loop that is the principal structural element interacting with EphA2 (Fig 3). The conserved glutamic acid residue within the GHephrin-A1 loop (E119ephrin-A1) forms a salt bridge with the conserved R103 in the GHEphA2 loop (Figs 3 and S3). The R103EphA2 is the most important residue for ephrin binding, as its mutation to glutamic residue entirely abolished the interaction [43]. We carried out comparative analyses of the gH/gL-EphA2 and ephrin-A1-EphA2 complexes further demonstrating that the structural elements employed by gH and gL resemble the ephrin-A1 ligand mode of binding to the EphA2 receptor. The GHephrin-A1 loop, which is the principal interaction region with the EphA2 receptor, occupies the same space as the gL “tail,” while the salt bridge established between the conserved R103EphA2 and E119ephrin-A1 is replaced at the same location by a salt bridge between R103EphA2 and E52gH (Fig 3). The conserved E52gH and E119ephrin-A1 occupy equivalent position in respect to R103EphA2, but the chain segments carrying the conserved E52gH and E119ephrin-A1 run in opposite directions so that the following residues, F53gH and F120ephrin-A1, do not superpose. Both F53gH and F120ephrin-A1 are engaged in π–π stacking interactions with F108ephrin-A1 and H53gL, respectively; in addition, the F108EphA2 establishes π–π interactions with Y21gL or F111ephrin-A1, indicating a common mechanism for stabilization of the GHephrin-A1 loop that presents the critically important glutamic acid residue for interactions with EphA2. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Structural mimicry between HHV-8 gH/gL and ephrin ligands. The EphA2 LBD from the EphA2 LBD–ephrin-A1 complex structure (PDB: 3HEI) was superimposed onto the EphA2 LBD from our complex. The same coloring scheme for gH/gL and EphA2 is applied as in Fig 1B, with the GHephrin-A1 loop highlighted in pink. For clarity, only the elements participating in the interactions are shown. The E119ephrin-A1 is indicated. Sequence alignment of a GH loop segment of ephrin-A ligands and the HHV-8 gH sequence are displayed to highlight the conservation of the glutamic acid that forms SBs with the EphA2R103 (E52gH and E119ephrin-A1). ephrin, Eph family receptor interacting protein; gH/gL, glycoproteins H and L; HHV-8, human herpesvirus 8; LBD, ligand-binding domain. https://doi.org/10.1371/journal.pbio.3001392.g003

Residues E52gH and R103EphA2 are critical for EphA2 dimerization on cells To test the importance of residue E52gH for binding of gH/gL to EphA2 in native membranes, FSI-FRET experiments were also performed with the E52RgH/gL recombinant protein, which exhibited significantly reduced binding to the soluble EphA2 ectodomains in BLI experiments (Fig 2D). The dimerization curve for EphA2 WT in the presence of E52RgH/gL is shown in Fig 5A and the fit parameters in Table 1. Constitutive EphA2 receptor dimerization was not observed. Rather, EphA2 interactions were reduced to levels similar to the case of no ligand (untreated), indicating that the presence of E52RgH/gL did not result in EphA2 dimer stabilization. This is consistent with the finding that the binding of this E52RgH/gL variant to EphA2 was disrupted and/or with the idea that bound E52RgH/gL did not enhance dimer stability. However, the measured Intrinsic FRET was slightly increased, as compared to no treatment, suggesting a decrease in the separation between the attached fluorescent proteins, and thus between the C termini of EphA2. This effect could be due to structural perturbations in the EphA2 dimer in response to possible E52RgH/gL binding at the high E52RgH/gL (200 nM) concentrations used, which could have propagated to the intracellular domain of EphA2. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. The gH/gL-induced EphA2 dimers on cells engage the “dimerization” interface. Dimerization curves calculated from the FSI-FRET data for (A) EphA2 WT in the presence of 200 nM gH E52R/gL mutant with mutation in EphA2 binding, and for the EphA2 mutants (B) R103EEphA2 mutant impaired in ligand binding, (C) G131YEphA2 mutant with mutation in DIN, and (D) L223R/L254R/V255REphA2 mutant with mutations in CIN. The data in A are compared to EphA2 WT data in the absence of ligand (untreated). The data in B–D were collected in the presence of 200 nM gH/gL and are compared to EphA2 WT in the presence of gH/gL (S10 Fig). No difference in the dimerization curve is observed with the mutated gHE52R/gL and thus does not induce constitutive EphA2 dimers as gH/gL does, which suggests impaired binding to EphA2. Large differences in the dimerization curves are observed for the R103EEphA2 and G131YEphA2 mutants, but the effect of the triple L223R/L254R/V255REphA2 mutation is modest. These data indicate that gH/gL-bound EphA2 dimers interact mainly via the DIN (where G131 EphA2 is engaged) but not via the CIN (where L223/L254/V255 EphA2 are engaged) and that R103EphA2 is important for gH/gL binding. The underlying data for all the panels can be found in S1 Data. CIN, clustering surface; DIN, dimerization interface; FSI-FRET, Fully Quantified Spectral Imaging–Förster Resonance Energy Transfer; gH/gL, glycoproteins H and L; WT, wild-type. https://doi.org/10.1371/journal.pbio.3001392.g005 In addition, we sought to test the importance of residue R103EphA2 for binding to gH/gL using the FSI-FRET method. The dimerization curve when the cells were transfected with EphA2 harboring the R103EEphA2 mutation in the gH/gL binding site, in the presence of saturating gH/gL concentrations, is shown in Fig 5B, and the fit parameters are shown in Table 1. The dimerization propensity for the R103EEphA2 variant in the presence of gH/gL is the same as for EphA2 in the absence of ligand (Table 1), indicating that either gH/gL binding to the R103EEphA2 mutant is disrupted, as also seen in the BLI experiments, and/or that binding did not lead to dimer stabilization. These data further corroborate our findings that R103EphA2 plays an essential role in gH/gL binding. Here again, we observed an increase in the Intrinsic FRET, which indicates that the fluorescent proteins are in closer proximity, as compared to EphA2 WT in the absence of ligand (Table 1). Similar to the behavior of the E52RgH /gL variant, this effect could be a consequence of R103AEphA2 binding to gH/gL, at the high gH/gL concentrations used, that would be transmitted to the EphA2 intracellular domains.

HHV-8 gH/gL induced EphA2 dimers on cell surface are stabilized via the “dimerization” interface We showed that when bound to gH/gL, EphA2 is a constitutive dimer. To determine if the EphA2 dimers form via one of the already described interaction surfaces, the dimerization (DIN) or clustering interface (CIN), as reported previously [31] (S2 and S3 Figs), we transfected HEK293T cells with the EphA2 variants with perturbed DIN (G131YEphA2) or CIN (L223R/L254R/V255REphA2) interfaces and treated them with soluble gH/gL ectodomains. The binding of these variants to gH/gL in solution, as measured by BLI, was not affected by the mutations, all of which reside outside of the gH/gL binding site (S3 Fig). The dimerization curves and FRET efficiencies for these mutants in the presence of gH/gL are shown in Fig 5C and 5D, respectively, with the fit parameters listed in Table 1. We observed a significant decrease in the dimerization due to the G131YEphA2 mutation (Table 1; Fig 5C), and a small effect due to the L223R/L254R/V255REphA2 mutations (Table 1; Fig 5D). Thus, G131EphA2 plays an important role in the stabilization of EphA2 dimers bound to gH/gL, implying that the EphA2 dimers are formed via DIN. This also supports our finding that gH/gL induces EphA2 dimers and not higher order oligomers, as these oligomers are known to engage both the DIN and the CIN (S2 Fig).

[END]

[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001392

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.