Sequence alignment shows that Met⁹³ and Lys²¹¹ are conserved in the NC1 domains of collagen IV. A multiple sequence alignment between the NC1 domains of bovine collagen IV α1-α6 chains shows that amino acids Met⁹³ (red) and Lys²¹¹ (blue) are conserved in all six chains. The red box around Lys²¹¹ denotes the consensus sequence for lysyl hydroxylase (53). The alignment was produced by ClustalW.

Isolation of α3α5-heterodimer and α4α4-homodimer. A, separation of NC1 hexamer subunits by a C₁₈ reverse-phase HPLC column. Protein elution was monitored by measuring absorbance at 214 nm. As shown in the chromatographic profile, fractions were grouped into pools P1 and P2. B, P1 and P2 fractions were analyzed by Western blots using antibodies Mab1 (anti- α1), Mab3 (anti-α3), H42 (anti-α4), and H52 (anti-α5) against NC1 domains. The migration of NC1 monomers and dimers is indicated by m and d, respectively. Fraction P1, composed mainly of α3α5-heterodimer, and fraction P2, enriched in α4α4-homodimer, were subsequently used in cross-link analyses.

Crypticity of GP epitopes in TBM NC1 hexamers. GP autoantibody binding to TBM NC1 hexamers was assayed by ELISA using three GP patient sera (GP1, GP2, and GP3) and a normal serum as control (CTL). GP sera were assayed for their binding to immobilized TBM NC1 hexamers (500 ng/well) under native (black bars) and dissociated (gray bars) conditions (GdnHCl). The assay was carried out as described under “Materials and Methods,” and the absorbance at 405 nm of triplicates was averaged. The TBM NC1 hexamers are a mixture of two populations, α3α4α5 and α1α2α1 NC1 hexamers, as described in Fig. 2.

Location of Met-Hyl and Met-Lys cross-links and quaternary structure of α3α4α5 D-hexamers. This model was generated by homology modeling using the coordinates of the α1α2α1 NC1 hexamer (22) and the cross-linking information obtained here. The NC1 hexamer is composed of two trimeric caps each consisting of α3(red), α4(blue), and α5(green) NC1 domain. The two heterotrimers interact through a large planar interface forming NC1 hexamer. Heterodimers (α3α5 and α5α3) and homodimers (α4α4), which are stabilized by Met-Hyl and Met-Lys cross-links (khaki color), respectively, were assembled into a NC1 hexamer. Molecular graphics images were produced using the University of California, San Francisco, Chimera (28).

A three-dimensional model for the D-isoform of the α3α4α5 NC1 hexamer showing the location and quaternary structure of the E_A and E_B epitopes. A solid surface representation (solvent exclusion) is shown for the three-dimensional atomic structure model for the α3α4α5 D-hexamer (center). The model depicts the location of the immunodominant epitopes for Goodpasture antibodies, E_A (yellow) and E_B (orange), within the α3 NC1 domain as well as the location of the S-hydroxylysyl-methionine and S-lysyl-methionine cross-links (khaki color). The left and right panels show portions of the α3-α5(left) and α5-α4(right) NC1 domain interfaces, which illustrate the structural setting of the E_A and E_B epitopes within the quaternary structure of the α3α4α5 D-hexamer. In each case, the NC1 hexamer was rotated about 120° in the direction marked by the arrows for optimal view. Lower insets represent the same view except that the α5(left) or α4(right) NC1 monomers were removed leaving a footprint (dotted line) on α3 NC1 monomer to visualize the residues that were sequestered by the neighboring α5 or α4 NC1 domain. Sequestered residues are labeled in white captions. Molecular surfaces images were produced using University of California, San Francisco, Chimera (28).

Experimental strategy for cross-linking analysis of NC1 dimers of the α3α4α5 D-hexamer. A shows that the α3α4α5 D-hexamer putatively contains two heterodimers, composed of α3(red) and α5(green) NC1 domains, and one homodimer, composed of α4(blue) NC1 domains. The NC1 dimers are presumed to contain Met-Hyl cross-links (black bars) that were previously found in theα1α1-homodimer (21). The Met-Hyl cross-link is formed between Met⁹³ from an NC1 domain of one trimer and a Hyl²¹¹ from an NC1 domain of the opposite trimer (Lys²¹¹ is post-translationally modified to Hyl²¹¹). A total of six Met-Hyl cross-links (two per NC1 dimer) can be formed at the trimer-trimer interface (21, 35). Dissociation of α3α4α5 D-hexamers yields α3α5-heterodimers and α4α4-homodimer. Trypsin digestion ofα3α5-heterodimers releases two peptide complexes (T-4481 and T-3959), each containing the Met-Hyl cross-link. Trypsin digestion ofα4α4-homodimer yields one complex (T-6498) containing the Met-Hyl cross-link, which is further digested with Asp-N endopeptidase to reduce its size (TA-3063 complex). B, proposed structure of the Met-Hyl cross-link based on mass spectrometry analysis of theα1α1-homodimer of placenta basement membrane (21). The side chains of Met⁹³ and Hyl²¹¹ are connected by a sulfonium ion. Upon fragmentation of the cross-link by CID, the covalent bond between C-γ and S-δ breaks, resulting in the loss of 48 atomic mass units (thiomethyl group) from Met⁹³ with a concomitant gain of 46 atomic mass units by Hyl²¹¹.

Characterization of the TBM NC1 hexamers with respect to NC1 domain composition and crypticity. A, TBM NC1 hexamers (lane 1) were fractionated using Mab1 (lane 2) or Mab3 (lane 3) monoclonal antibodies to NC1 hexamers containingα1 andα3 subunits, respectively. Mab1- and Mab3-bound fractions were precipitated with protein G beads, resolved by SDS-PAGE into NC1 monomer (m) and NC1 dimer (d) subunits, and then analyzed by Western blot in five replica membranes for the presence of α1-α5 NC1 domains, using monoclonal antibodies H1 (anti-α1), H22 (anti-α2), H31 (anti-α3), H43 (anti-α4), and H53 (anti-α5). Lane 1 shows the staining of TBM NC1 hexamer (total) which serves as reference for the migration of NC1 domain subunits. B, diagram of the immunoprecipitation (IP) with Mab1 and Mab3 separating the two distinct NC1 hexamer populations. C, immunoprecipitation of NC1 domains from dissociated (lane 2) or native (lane 3) TBM NC1 hexamers using GP antibodies (Ab) (IgG fraction). GP-bound NC1 hexamers were subjected to Western blotting analysis for the presence of α1-α5 NC1 domains as described for Fig. 2A. TBM NC1 hexamers were dissociated (Dissoc) with 6 m GdnHCl for 15 min at 60 °C. D, diagram illustrating the immunoprecipitation of the different forms of α3 NC1 domains from native and dissociated NC1 hexamers.

Mass spectrometry analysis of tryptic peptide complexes derived from α3α5-heterodimer and α4α4-homodimers. A, fraction P1 (Fig. 3) containing mainly α3α5-heterodimers was digested with trypsin and analyzed by LC-ESI/MS³. Left panel, ESI spectrum of the T-4481 complex constituted by NDYSYWLSTPEPMPMSMQPLK (T-2514) and KPIPSTVK_OHAGELENIISR (T-1967) peptides. Multiply charged ions of the complex (+3, +4, +5, and +6) are indicated in red for better visualization. Right panel, MS/MS spectrum generated by fragmentation of the quintuply charged ion 897.8 m/z produces the 823.8 m/z and 1007.87 m/z ions. The 823.8 m/z is the triply charged ion for T^*-2466 peptide, which represents the loss of 48 atomic mass units by the T-2514 peptide. The 1007.87 m/z is the doubly charged ion for T^*-2013 peptide, which represents the gain of 46 atomic mass units by the T-1967 peptide. The asterisk indicates that such tryptic peptide contains a modified residue. The sequences of T^*-2466 and T^*-2013 peptides, including their amino acid modifications, were confirmed by further fragmentation (MS³) of 823.8 m/z and 1007.87 m/z ion, respectively (supplemental Fig. 9A). B, left panel, ESI spectrum of the T-3959 complex constituted by NDYSYWLSTPAMIPMDMAPITGR (T-2629) and PQSETLK_OHAGDLR (T-1330) peptides. Multiply charged ions of the complex (+3, +4, and +5) are indicated in red for better visualization. Right panel, MS/MS spectrum generated by fragmentation of the 990.76 m/z ion produces the 1291.91 m/z and 688.90 m/z ions. The 1291.91 m/z is the doubly charged ion for T^*-2581 peptide, which represents the loss of 48 atomic mass units by the T-2629 peptide. The 688.90 m/z is the doubly charged ion for the T^*-1376 peptide, which represents the gain of 46 atomic mass units by the T-1330 peptide. The asterisk indicates that such tryptic peptide contains a modified residue. The sequence of T^*-2581 and T^*-1376 peptides, including their amino acid modifications, were confirmed by further fragmentation (MS³) of 1291.91 m/z and 688.90 m/z ions, respectively (supplemental Fig. 9B). C, LC-ESI/MS³ analysis of fraction P2 (Fig. 3), containing mainly α4α4-homodimers, after digestion with trypsin and Asp-N endoproteinase. Left panel, ESI spectrum of the TA-3665 complex constituted by STAPLPMTPLSEDEIRPY (TA-2016) and SAPLPDTLKESHAQR (TA-1649) peptides. Multiply charged ions of the complex (+3, +4, +5, and +6) are indicated in red for better visualization. Right panel, MS/MS spectrum generated by fragmentation of 917.1 m/z ion produces the 985.44 m/z and 848.60 m/z ions. The 985.44 m/z is the doubly charged ion for TA^*-1968 peptide, which represent the loss of 48 atomic mass units by TA-2016 peptide. The 848.60 m/z is the doubly charged ion for the TA^*-1695 peptide, which represents the gain of 46 atomic mass units by the TA-1649 peptide. The asterisk indicates that such peptide contains a modified residue. The sequences of TA^*-1968 and TA^*-1695 peptides, including their amino acid modifications, were confirmed by further fragmentation (MS³) of 985.44 m/z and 848.60 m/z ions, respectively (supplemental Fig. 9C).