Direct interaction of the mouse cytomegalovirus m152/gp40 immunoevasin with RAE-1 isoforms

Abstract

Cytomegaloviruses (CMVs) are ubiquitous species-specific viruses that establish acute, persistent, and latent infections. Both human and mouse CMVs encode proteins that inhibit the activation of natural killer (NK) cells by downregulating cellular ligands for the NK cell activating receptor, NKG2D. The MCMV glycoprotein m152/gp40 downregulates the surface expression of RAE-1 to prevent NK cell control in vivo. So far, it is unclear if there is a direct interaction between m152 and RAE-1 and, if so, if m152 interacts differentially with the five identified RAE-1 isoforms, which are expressed as two groups in MCMV-susceptible or -resistant mouse strains. To address these questions, we expressed and purified the extracellular domains of RAE-1 and m152 and performed size exclusion chromatography binding assays as well as analytical ultracentrifugation and isothermal titration calorimetry to characterize these interactions quantitatively. We further evaluated the role of full-length and naturally glycosylated m152 and RAE-1 in cotransfected HEK293T cells. Our results confirmed that m152 binds RAE-1 directly, relatively tightly (K(d) < 5 microM), and with 1:1 stoichiometry. The binding is quantitatively different depending on particular RAE-1 isoforms, corresponding to the susceptibility to downregulation by m152. A PLWY motif found in RAE-1beta, although contributing to its affinity for m152, does not influence the affinity of RAE-1gamma or RAE-1delta, suggesting that other differences contribute to the RAE-1-m152 interaction. Molecular modeling of the different RAE-1 isoforms suggests a potential site for the m152 interaction.

Figure 1. m152 binds to RAE-1 directly

A,B, Purification of RAE-1 and m152. The ectodomains of RAE-1 isoforms and of m152 were expressed and purified as described in Materials and Methods. Shown here are the chromatographs of RAE-1β (A, pink) and m152 (B, green) overlaid with the protein molecular weight standards (dashed line). C,D,E, Binding of m152 to RAE-1β, γ, and δ examined by a SEC binding assay. m152 and RAE-1 were mixed in equimolar amounts, incubated for 30 min on ice, analyzed chromatographically (blue), and compared to the profile of individual m152 (green) and RAE-1 proteins (pink) alone. Insets show enlargements of the peak regions. (Flow rate was 0.75 ml/min, so elution volumes may be calculated from the indicated time of elution.)

Figure 2. Biophysical experiments show m152 binds to RAE-1 at 1:1 ratio and with micromolar K_d

A, Sedimentation coefficient distribution c(s) measured in sedimentation velocity (SV) experiments for 11.3 μM RAE-1 alone (dotted line) and 11.3 μM m152 alone (thin line), and 0.63:1 mixtures of RAE-1 and m152 at 1.2 μM (thick line), 4.5 μM (open triangles), and 11.3 μM (filled triangles). Integration of c(s) traces over the range from ~1.5 to ~4.5 S or ~2.5 to ~4.5 S (indicated by horizontal arrows) yields values for the weighted-average s-value s_w or the reaction boundary s-value s_fast, respectively. B, Isotherms of s_w (filled symbols) and s_fast (open symbols) as a function of m152 concentration from dilution series at molar ratios RAE-1: m152 of 1.25:1 (black), 0.63:1 (green), and 2.5:1 (blue). The solid lines shown are from a global fit with a 1:1 binding model that incorporated additional data from the relative amplitude of the reaction boundary, as well as the data from isothermal titration calorimetry (ITC). The best-fit s_w value for the complex was 4.3 S, and the binding constant K_d = 4.1 μM. C, Equilibrium absorbance profiles at 280 nm of the mixture of 2.9 μM m152 and 10.2 μM RAE-1 sedimenting at 11,000 (circles) and 17,000 rpm (triangles), their best-fit distribution (solid lines) and the signal predicted for the complex (dotted and dashed lines). These data are taken from a global fit with a 1:1 binding model of 20 equilibrium gradients at different concentrations, rotor speeds, and absorbance wavelengths. D, Enthalpy isotherm from the titration of 100 μM RAE-1 into 10 μM m152 in ITC (symbols), based on a single titration series. The solid line shows the best-fit isotherm from a global fit with the SV isotherm data.

Figure 3. The PLWY motif contributes to the binding affinity of RAE-1β to m152

A, SEC binding assay. The PLWY motif was deleted from RAE-1β and γ and inserted into RAE-1δ by site-directed mutagenesis. The binding of wild type and mutant RAE-1 proteins to m152 was examined by SEC as described in Materials and Methods. The results from wild type RAE-1 proteins are the same as shown in Figure 1C, D, E and shown here for ease of comparison. B, Sedimentation velocity ultracentrifugation. Isotherms of weighted-average sedimentation coefficients s_w (in experimental units) obtained from SV experiments of equimolar mixtures of RAE-1 and m152 (symbols) and best-fit isotherm with a 1:1 binding model (solid lines) using the pre-determined s-value of RAE-1 and m152, and a complex s-value fixed to that determined in the extensive characterization of RAE-1β/ m152 interaction of Figure 2. The insets show the sedimentation coefficient distributions c(s) underlying the s_w data points, for approximately equimolar loading concentrations of 1 (cyan), 2 (magenta), 7 (blue), 14 (green), and 32 (red) μM. Precise concentrations were measured from the interference optical signal and varied slightly.

FIGURE 4. RAE-1 isoforms differ in PLWY motif as well as in charged residues

Amino acid sequence alignment of RAE-1 isoforms. The amino acid sequences of the extracellular domain of the five RAE-1 isoforms were aligned using ClustalW (45). Secondary structure elements of RAE-1β (PDB: 1JFM, (26)) are shown above the alignment. The PLWY motif is boxed in blue. Cysteine residues involved in the formation of disulfide bonds are shown as green dashed line. Charged residues involved in the formation of potential salt bridges are shown as blue dashed line. Differences in charged residues among RAE-1 isoforms are highlighted as black triangles. Predicted N-linked glycosylation sites are indicated as cyan ovals. The figure was generated with ESPRIPT 2.2 (46).

Figure5. m152 differently downregulates surface expression of RAE-1 isoforms

A, A constant amount of wild type and the PLWY mutant of RAE-1γ or δ cDNAs were cotransfected into human HEK293T cells along with an empty vector carrying an IRES-GFP (left and middle column) or the corresponding m152-encoding vector (right column) at increasing amounts as indicated. 48 hours after transfection, cells were harvested and stained with either an isotype control Ig (left column) or rat pan anti-RAE-1 mAb (186107, middle and right column) and analyzed for surface RAE-1 expression by flow cytometry. Results are representative of three independent experiments. Shown here is the two-dimensional dot plot of RAE-1γ and m152 cotransfection experiment. B, Summary of the above described cotransfection experiments in histogram. Red, isotype control; blue, cotransfected RAE-1 with empty vector and stained with 186107; green, cotransfected RAE-1 with m152 and stained with 186107.

Figure 6. RAE-1 isoforms differ in surface electrostatic charge

A, Models of RAE-1γ and δ were constructed in Coot based on the crystal structure of RAE-1β (PDB 1JFM chain A (26)). RAE-1β shows 92 and 94% amino acid sequence identity with RAE-1γ and δ, respectively. Residues in RAE-1β were replaced with the corresponding residues of RAE-1γ and δ followed by geometry idealization as described in Material and Methods. The molecular surfaces are colored by electrostatic potential, with positively charged areas in blue and negatively charged areas in red. B, Ribbon diagrams of RAE-1β, γ, and δ. Positions of some of the polymorphic charged residues are shown as stick representations superposed on the ribbons. For RAE-1β, residues Lys89 (blue), Glu178 (red), and Lys181 (blue) are indicated; for RAE-1γ, Glu89 (red), Asp97 (red), and Asp178 (red); and for RAE-1δ, Glu89 (red), Asp97 (red), Glu178(red), and Lys181(blue) are shown. (Numbering is based on the alignment in Figure 4).