Key to Collagen Cross-Links
By: Carol A. Rouzer, VICB Communications
Published: August 28, 2012
The discovery of peroxidasin as the enzyme that cross-links basement membrane collagen reveals an important role for hypohalous acid.
The epithelial cells that line body cavities, and the endothelial cells that line blood vessels are supported by a thin network of fibers known as the basement membrane. This membrane not only provides structural integrity to the epithelium and endothelium, it also serves as a scaffold for the assembly of macromolecules and interacts with the surface receptors of cells in the substructure of the tissue. The single major constituent of the basement membrane is collagen IV, which is constructed from a triple helical protomer as its basic building block. Unlike the collagen used to form most connective tissues, such as tendons and ligaments, collagen IV polypeptides retain their C-terminal non-collagenous domain (NC1), which forms an expanded, nonhelical structure at the end of the protomer. To form the network of fibers that make up the basement membrane, the NC1 regions of two protomers aggregate end-to end to yield a hexamer (Figure 1). This hexameric structure is then stabilized by the formation of a sulfilimine bond (S=N) cross-link between methionine-93 (Met93) of one protomer and hydroxylysine 211 (Hyl211) of the opposing protomer. Sulfilimine bond formation is unique to collagen IV, yet the responsible metabolic pathway for its synthesis has not been identified. Now, Vanderbilt Institute of Chemical Biology investigator Billy Hudson and his laboratory along with collaborators in John Fessler’s lab at UCLA have solved the mystery [G. Bhave et al., (2012) Nat. Chem. Biol., published online July 29, DOI:10.1038/nchembio.1038].
Figure 1. Cartoon representation of the molecular structure of the NC1 domains of two collagen trimers joined end-to-end to form a hexamer. Sulfilimine bonds between Met-93 of one monomer and Hyl211 of the opposing monomer stabilize the structure. Image reproduced through the courtesy of Wikimedia Commons under the GNU Free Documentation License.
The Hudson lab used the PFHR-9 mouse endodermal cell line as a source of collagen IV. These cells produce substantial amounts of the protein, and confluent cultures also form mature basement membrane. Digestion of the collagen IV produced by PFHR-9 cells with collagenase allowed the investigators to isolate NC1 hexamers with intact sulfilimine bonds confirmed by mass spectral analysis. Because prior investigation suggested that a matrix oxidase could be involved in sulfilimine bond formation, the Hudson lab examined the effect of three distinct peroxidase inhibitors, phloroglucinol, methimazole, and 3-aminotriazole on cross-link formation. All three inhibitors reduced the number of sulfilimine bonds in collagen IV produced by the cells without affecting collagen IV hexamer assembly or breaking pre-exisiting sulfilimine cross-links.
Incubation of confluent PFHR-9 cells in the presence of peroxidase inhibitors led to the production of basement membrane lacking sulfilimine cross-links. The investigators found that incubation of this material with H2O2 in the absence of peroxidase inhibitors resulted in efficient cross-link formation. This did not occur in basement membrane that had been pretreated with guanidinium chloride to denature intrinsic enzymes that might be present. These results suggested that a peroxidase present in the basement membrane catalyzed sulfilimine bond formation.
Knowing that N3- (azide ion) inactivates most peroxidases by reacting with the active site heme to form an organic azide, the Hudson lab treated isolated basement membrane with H2O2 and azide. They then solubilized the proteins and used “click” chemistry to couple the organic azide with alkyne-labeled biotin. Analysis of the resulting sample by western blot using streptavidin-linked horseradish peroxidase revealed a single major protein band of 160 to 200 kDa molecular mass. Streptavadin agarose affinity chromatography provided the purified protein, which was then identified by mass spectrometric analysis to be the well known enzyme, peroxidasin.
Purified recombinant human peroxidasin, in the presence of H2O2, readily catalyzed cross-link formation of NC1 hexamers in vitro. Similar results using the enzyme from Drosophila melanogaster confirmed the enzyme’s role across species. Knowing that many peroxidases use halide ions as part of their reaction mechanism, the Hudson lab investigated peroxidasin’s requirement for Cl- or Br- ion. They found that very little sulfilimine bond formation could occur in the absence of halide. This result suggested that peroxidasin-catalyzed sulfilimine bond formation likely occurs through the action of a HOBr or HOCl intermediate, which reacts with the sulfur of Met93 to form a halosulfonium ion. Trapping of this intermediate by the amine nitrogen of Hyl211 forms the sulfilimine bond (Figure 2). Support for this mechanism came from the finding that incubation of NC1 hexamers with HOBr or HOCl in the absence of enzyme led to cross-link formation. In fact, other peroxidases, such as myleoperoxidase and eosinophil peroxidase, which convert halide ions to their corresponding hypohalous acids, could catalyze NC1 hexamer cross-link formation, while lactoperoxidase, which forms hypohalous acids poorly, did not.
Figure 2. Proposed mechanism for sulfilimine bond formation catalyzed by peroxidasin. The enzyme first generates hypohalous acid following reduction of H2O2. Reaction of the hypohalous acid with the sulfur of Met93 yields a halosulfonium intermediate, which then reacts with the amine nitrogen of Hyl211 to form the sulfilimine bond. Reprinted with permission from Macmillan Publishers Ltd. from G. Bhave et al., (2012) Nat. Chem. Biol., published online July 29, DOI:10.1038/nchembio.1038. Copyright 2012.
The ability of multiple peroxidases to catalyze sulfilimine bond formation in vitro led the investigators to test their activities within the basement membrane structure. They used confluent PFHR-9 cells grown in the presence of peroxidase inhibitors to generate basement membrane lacking cross-links. They then plated HEK293 cells expressing peroxidasin on the basement membrane and demonstrated that cross-link formation had occurred. In contrast, cells expressing myeloperoxidase or lactoperoxidase could not catalyze cross-link formation. These findings suggest that only peroxidasin is capable of generating hypohalous acid and using it to form sulfilimine cross-links within the complex basement membrane structure.
To explore the importance of peroxidasin to sulfilimine bond biosynthesis in vivo, the Hudson group worked with the Fessler lab to explore basement membrane formation in the larvae of D. melanogaster that express a hypomorphic peroxidasin. This fatal mutation leads to death at the third instar larval stage. The investigators found that collagen networks in the basement membranes of the midgut in the larvae were severely distorted and badly torn (Figure 3). Analysis of the collagen IV from the larvae revealed a marked decrease in the content of cross-linked monomers.
Figure 3. Comparison of the basement membrane collagen network from wild-type D. melanogaster larvae (Pxn+/+) and larvae heterozygous (Pxn+/-) and homozygous (Pxn-/-) for a hypomorphic peroxidasin. Note the tears and large holes in the fiber matrix from Pxn-/- larvae. Reprinted with permission from Macmillan Publishers Ltd. from G. Bhave et al., (2012) Nat. Chem. Biol., published online July 29, DOI:10.1038/nchembio.1038. Copyright 2012.
These findings answer an important question concerning the basic biochemistry of basement membrane collagen formation. The investigators note that they may also have important implications for the pathogenesis of a number of disease states. The generation of hypohalous acid at low concentrations for sulfilimine bond formation is obviously of great benefit. However, under conditions of inflammation or oxidative stress, when H2O2 levels are abnormally increased, it is possible that basement membrane peroxidasin could produce toxic levels of hypohalous acid, leading to tissue damage. In this regard, it is noteworthy that peroxidase expression is increased in diseases such as hypertension and atherosclerosis. Under these conditions, excessive cross-link formation and hypohalous acid-mediated damage could contribute to the fibrosis seen in these diseases.