Proteomic characterization of type I collagen N-terminal crosslinked peptides

Abstract

Collagen cross-links mediated by the lysyl oxidase and lysyl hydroxylase families of enzymes significantly contribute to the biomechanical strength and rigidity of tissues, influencing cell signaling and the downstream cell phenotype. In the clinic, the proteolytically liberated N-terminal cross-linked peptide of collagen I (NTX) is used as a biomarker of bone and connective tissue turnover, which is altered in several disease processes. Despite the clinical utility of these collagen breakdown products, the majority of the cross-linked peptide species have not been identified in proteomic datasets. Here, we evaluate several parameters for the preparation and identification of these peptides from the collagen I-rich Achilles tendon. Our refined approach, which involves chemical digestion for protein solubilization coupled with mass spectrometry, enables the identification of NTX cross-links in a range of modification states. We then applied a spectral library approach to identify differences in collagen cross-links in bovine pulmonary hypertension. The presented method offers unique opportunities to understand extracellular matrix remodeling events in development, aging, wound healing, and fibrotic disease that modulate collagen architecture through lysyl hydroxylase and lysyl oxidase enzymes.

Keywords: Collagen; Cross-linking; Lysyl oxidase; Mass spectrometry; Proteomics; Spectral library; Telopeptides.

Fig. 1

Lysyl oxidase (LOX) and lysyl hydroxylase (LH) mediated cross-linking at aligned sites within the distinct collagen fibril banding pattern. A.) Overview of collagen fibril organization. B) LH and LOX modification of lysines followed by spontaneous cross-linking to form HLKNL product. Reduction to DHLNL is also shown. C.) Generation of trivalent PYD cross-links D.) D-banding pattern observed in collagen fibrils and corresponding lysine alignment across the 5 registries (M1-M5). Reported sites for COL1 cross-linking are labeled in bold: N-terminal (NT), C-terminal (CT), N-helix (NH), C-helix (CH). N-telopeptides from M1′ align with the M5 OL1A1 site K1107 (K946 mature numbering) and COL1A2 K1021 (K942). These have historically been referred to as K930 and K933, respectively [7,[10], [11], [7], [12], [13]].

Fig. 2

Proteomic approach used for NTX peptide characterization. A.) Schematic of the analytical workflow used to provide multiple lines of evidence for identified NTX species (further detail can be found in Supplemental Table 3). B.) Global composition of the bovine Achilles tendon sample used for method development.

Fig. 3

MS/MS spectra of the COL1A1-COL1A2 NTX dimer peptides cross-linked by the HLKNL linkage. COL1A1-derived fragment ions are highlighted in red, and COL1A2-derived fragments are in blue. Fragment ions in black can be assigned to both sequences in the cross-linked peptide. Low-intensity peaks corresponding to additional fragment ions are not labeled here. Modified residues: N-terminal pyroglutamic acid (q), oxidized Met (m), hydroxyproline (p), homoserine (m’), and homoserine lactone (m’’). Fragment modifications: neutral loss of SOCH₄ from oxidized methionine sidechain (−64) and water loss (b₀). A) HCD spectrum of precursor (pr) at m/z 1095.495 (4 + ). B) ETD spectra of m/z 1095.495(4 + ), representing the NTX dimer identified from the hydroxylamine-treated sample. C) HCD spectrum of m/z 1030.800(3 + ), representing the NTX dimer identified from the CNBr-treated sample. HCD spectra were acquired in the Orbitrap analyzer, and ETD spectra in the linear ion trap. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Trivalent cross-linked NTX peptides with the COL1A1CH peptide. A) HCD spectrum of m/z 1005.263(5 + ) representing the α1NT-α2NT-α1CH_PYD trimer peptide identified via HA digestion. PYD = hydroxylysyl-pyridinoline linkage. The α1NT-related fragment ions are highlighted in red and α2NT-related fragments are in blue. Fragment ions in black can be assigned to both sequences. No specific fragment ions for the helical COL1A1 sequence were detected; its presence was only confirmed by the additive mass of the peptide. Modifications: N-terminal pyroglutamic acid (q), oxidized Met (m), neutral loss of SOCH₄ from oxidized methionine sidechain (−64). HCD spectra acquired in the Orbitrap analyzer (A & C). B) ETD spectrum of m/z 1005.263(5 + ) representing the α1NT-α2NT-α1CH species cross-linked by the PYD linkage identified in an HA-treated sample. α1NT-derived fragment ions are highlighted in red, α2NT-derived fragments are in blue, and α1CH-derived fragments are in black. ETD spectrum acquired in the linear ion trap. C) HCD spectrum of m/z 1163.544(5 + ) representing the α1NT-α1NT-α1CH trimer peptide cross-linked by the PYD linkage identified in an HA-treated sample. α1NT-derived fragment ions are highlighted in red, and α1CH-derived fragments are in black. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5

LC-MS data highlighting the ion envelope landscape of large linear and cross-linked collagen peptides. SEC fractionated, HA then trypsin digested, bovine tendon MS1 spectra with identified peptides annotated. A) Chromatogram trace of SEC fractionation, with the fractions containing identified crosslinked species highlighted in yellow. B) Enlarged panel highlighting a region within F1. The shaded blue region represents the total LFQ-quantified area for α1NT-α1NT_HLKNL. The quantified proportion of the ion envelope is consistent across most cross-linked peptide identifications. C) MS1 trace of the ion envelope for α1NT-α2NT_HLKNL seen in F2. F1-3 represents the first three fractions collected. The regions outlined in blue represent the ion envelope for the annotated identification. D) F1 contains the majority of the α 1NT-α1NT- α 1CH trivalent cross-linked collagen peptide identifications. E) F2 contains most of the divalent cross-linked peptides and the trivalent α 1NT- α 2NT- α 1CH. F) F3 consists mostly of linear collagen peptides (∼1% of identified cross-link signal). Modifications: N-terminal pyroglutamic acid (q), oxidized Met (m), hydroxyproline (p). Blue labels are identified linear peptides; red are cross-linked identifications, and gold are the mature cross-links containing three peptides. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 6

Spectral library searching for identification and quantification of NTX cross-links in bovine pulmonary hypertension. A) Spectrum match for α1NT-α2NT_HLKNL_Ox identification from a bovine hPH model using an unfractionated bovine main pulmonary artery (mPA) sample (top identification). A spectral library (SL) searching approach was used by creating a spectral library from the highly fractionated sample set (lower identification). Fragment ions are red for y ions, blue for b ions, and black for non-y or b ions B) Extracted ion chromatograms for identification of α1NT-α2NT_HLKNL_Ox in unfractionated control (CO) and pulmonary hypertension (PH) mPA samples (m/z 1465.66, 3 + ). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 7

X-ray diffraction reveals unassigned density in position of collagen cross-links. A) Density map generated by x-ray fiber diffraction displaying the region of overlaid COL1A1 and COL1A2 chains containing unassigned inter-chain density [64] indicated with a yellow arrow (right-hand side, density displayed 5X compressed in the y-axis for display purposes). B) α1NT-α2NT-α1CH_PYD cross-link modeled into missing density between COL1 α 1NT (green, bottom molecule), COL1 α 2NT (blue, bottom molecule), and COL1 α 1CH (green, upper molecule). Additional COL1 α 1NT chains not cross-linked at this location are colored magenta. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)