Identification of a major epitope recognized by PLA2R autoantibodies in primary membranous nephropathy

Abstract

Phospholipase A2 receptor 1 (PLA2R) is a target autoantigen in 70% of patients with idiopathic membranous nephropathy. We describe the location of a major epitope in the N-terminal cysteine-rich ricin domain of PLA2R that is recognized by 90% of human anti-PLA2R autoantibodies. The epitope was sensitive to reduction and SDS denaturation in the isolated ricin domain and the larger fragment containing the ricin, fibronectin type II, first and second C-type lectin domains (CTLD). However, in nondenaturing conditions the epitope was protected against reduction in larger fragments, including the full-length extracellular region of PLA2R. To determine the composition of the epitope, we isolated immunoreactive tryptic fragments by Western blotting and analyzed them by mass spectrometry. The identified peptides were tested as inhibitors of autoantibody binding to PLA2R by surface plasmon resonance. Two peptides from the ricin domain showed strong inhibition, with a longer sequence covering both peptides (31-mer) producing 85% inhibition of autoantibody binding to PLA2R. Anti-PLA2R antibody directly bound this 31-mer peptide under nondenaturing conditions and binding was sensitive to reduction. Analysis of PLA2R and the PLA2R-anti-PLA2R complex using electron microscopy and homology-based representations allowed us to generate a structural model of this major epitope and its antibody binding site, which is independent of pH-induced conformational change in PLA2R. Identification of this major PLA2R epitope will enable further therapeutic advances for patients with idiopathic membranous nephropathy, including antibody inhibition therapy and immunoadsorption of circulating autoantibodies.

Keywords: Anti-PLA2R; PLA2R; autoantibody; epitope; membranous nephropathy.

Figure 1.

Epitope recognition in PLA2R fragments is dependent on denaturing/non-denaturing conditions. (A) Representations of the extracellular subdomains of a long-fragment PLA2R (N-C8) comprising the N-terminal ricin domain (or cysteine-rich domain), fibronectin type II domain, and 8 C-type lectin domains (CTLD), two shorter fragments ending after the first 3 C-type lectin domains (N-C3) and the first 2 CTLD (N-C2) and the ricin domain. Silver-stained SDS-PAGE gel of purified recombinant N-C8, N-C3, N-C2, and ricin proteins under nonreducing conditions; lane 1 is a molecular mass marker, and subsequent lanes show 1 µg of purified recombinant N-C8 (180 kD), N-C3 (90 kD), N-C2 (65 kD), and ricin (18 kD) proteins. (B) Western blotting analysis of denatured N-C8, N-C3, N-C2, and ricin proteins under nonreduced conditions using a pool of five human sera and rabbit antibodies to PLA2R. Slot blotting analysis of nondenatured N-C8, N-C3, N-C2, and ricin fragments under nonreduced and reduced conditions.

Figure 2.

Human anti-PLA2R shows similar binding affinity to N-C8, N-C3, N-C2 and the ricin domain. (A) Western blot analysis of fragmented N-C3 peptides by OFFGEL fractionation using different patient sera. (B) Representative sensorgrams derived from injections of different concentrations of affinity purified anti-PLA2R antibodies over immobilized N-C8 (left panel), N-C3 (right panel), and N-C2 and ricin domain (bottom panels). The tables summarize the kinetic constants of four separate runs for the binding of purified human PLA2R-specific antibody to the PLA2R fragments. Results were obtained after reference subtraction. Kinetics data were fitted to a Langmuir 1:1 interaction model. The association (k_a), dissociation (k_d), and equilibrium (K_D) constants for each run were similar and revealed an overall high binding affinity with an apparent K_D of approximately 0.1 nM.

Figure 3.

Human anti-PLA2R binding to N-C8 is unaffected by a pH induced conformational change. (A) pH-dependent conformational change of N-C8 determined by dual polarization interferometry. Changes in the thickness of the N-C8 layer as a function of decreasing pH was measured and represented the adsorption behavior of different folded states. (B) Sedimentation velocity of N-C8 and N-C3 (inset) at a pH of 6.2 (black line) and of 7.2 (gray line) analyzed by the distribution of Lamn equation solutions c(s) model using the program Sedfit. (C) Single-cycle kinetics between immobilized purified human PLA2R-specific antibody and N-C8 at a pH of 6.2 and of 7.2. A range of 1, 2, 5, 10, and 15 nM of N-C8 was injected, one cycle at a pH of 6.2 and the other at a pH of 7.2. The surface was not regenerated between injections in each cycle.

Figure 4.

The epitope in the cysteine rich ricin domain is composed of a 31-mer peptide. (A) Screening of inhibitory peptides identified by MS. Sequences of the synthesized peptides as well as their coverage is shown on the schematic diagram of the PLA2R N-C3 fragment. Nine peptides including a control were checked for their inhibitory activity against N-C3. Purified human PLA2R-specific antibody was incubated separately with an excess of each peptide (5 µM) then injected onto the immobilized N-C3. (B) Dose-response curves of inhibition between purified human anti-PLA2R and N-C3 by an increasing amount of peptide 2 and 31-mer peptide. Error bars indicate SDs. (C) Sensorgrams derived from injections of different concentrations of affinity purified PLA2R-specific antibody over captured 31-mer peptide. Kinetics data were fitted to a Langmuir 1:1 interaction model with mass transfer.

Figure 5.

Characterization of the binding pocket. (A) Slot blots analysis of the 31-mer peptide demonstrating the reduction in binding to the autoantibody under reducing conditions. 3D cartoon representation of the 31-mer peptide structure comprising peptide 1 and peptide 2 (dark gray) with the disulfide bond bridging the two peptides. Inset: gel filtration elution profiles of the nonreduced and reduced 31-mer peptide showing its cyclic nature in native form. (B) SPR inhibition analysis of the binding between the human autoantibody and captured N-C3 (left) or N-C2 (right) by the human 31-mer peptide.

Figure 6.

3D PLA2R model structure and arrangement of domains. (A) Representative area micrographs of N-C8 with highlighted images of single molecules within white squares, selected projections obtained and two-dimensional averages of the images within the corresponding class. Scale bar=1000 Å. TEM, transmission electron microscopy; MSA, multi statistical analysis. (B) Electron density of N-C8 solved from approximately 20,000 selected particles at 20-Å resolution and a 3D reconstruction of N-C8 showing the domains arrangement. Individual PLA2R domains were modeled using Phyre2. The domains were aligned in the electron microscopic density map using Chimera and validated using area under the curve hydrodynamics parameters (Table 2). Scale bar=50 Å.