HS-GAG degradation (R-HSA-2024096)

[ukemi]

• Create a mouse model
• Change mapping of R-HSA-2024096 from glycosaminoglycan catabolic process (GO:0006027) to heparan sulfate proteoglycan catabolic process (GO:0030200)
• Should this blurb "As part of the natural turnover of GAGs, extracellular KSPGs translocate to the lysosome to be degraded." on R-HSA-2024084 read "As part of the natural turnover of GAGs, extracellular HSPGs translocate to the lysosome to be degraded. " ? Yes - more - the reference cited for this statement describes glycoproteins as having either intra- or extra-cellular locations, so the route to the lysosome is ambiguous. The specific glycoproteins annotated here (as also for keratan sulfate catabolism) are associated with the plasma membrane or are extracellular, so they must travel by endocytosis (GO:0006897). So there are two fixes: add GO:0006897 as a GO BP attribute for reaction R-HSA-2024084, and fix its summation / blurb to "As part of the natural turnover of GAGs, extracellular HSPGs are endocytosed to the lysosome to be degraded."
• I think we need to reexamine the Reactome representation of HPSE2. It looks like it doesn't function as an enzyme, but rather inhibits HPSE (PMID:20576607). See figure 6. Also interesting PMID:34681753, 33585253, 30487472. Agreed. Reaction R-HSA-1678694 has been reconstructed to show HPSE:HS-proteoglycan binding at the cell surface with no cleavage and no uptake into the cell - see the new version here.
• I think it's reasonably clear that HPSE gets internalized by endocytosis. See the above papers. Let's do a sanity check on the mouse model: http://noctua.geneontology.org/editor/graph/gomodel:627d95ee00000275. And (done - item 2) add GO "endocytosis" BP term to Reactome reaction. But the model shown in PMID:20576607 Figure 6 is a process in which HPSE is secreted into the extracellular space where it forms a complex with HS-proteoglycan which is then endocytosed and targeted to the lysosome. As opposed to the Reactome process in which HPSE resides in the lysosome and HS-proteoglycan is delivered to it there. But the difference is a bit more subtle. HPSE could get to the lysosome in the first place (before the events shown in the pathway) by secretion and re-uptake, as many lysosomal enzymes do. But Figure 6 suggests more than that, that every round of HS-proteoglycan uptake for breakdown is accompanied by fresh HPSE uptake. See what you think of the evidence - I could re-do the Reactome pathway to correspond to Figure 6. Done? - see the newly added first three steps of HS-GAG degradation (R-HSA-2024096) on the Reactome internal website. I'm still not convinced by PMID:20576607 that a normal role for HSPE2:HS-syndecan complexes on the cell surface is to sequester and inhibit HSPE1.
• N-Acetylglucosamine-6-Sulfate Sulfatase (GNS) in the degradation PMID:99718. A quick search of pathway diagrams has this after beta-glucuronidase, implying that R-HSA-2090043 doesn't completely desulfonate the GAG. PD comment: we asserted that SGSH removes sulfates from terminal N-sulfoglucosamine residues, so I guess that it can't touch sulfates attached elsewhere on the oligosaccharide, but only ones that become terminal as the oligosaccharide is cleaved by other enzymes? This brings up another limitation of the Reactome annotation - there is a range of oligosaccharide lengths and compositions so we haven't really given a recipe for systematic catabolism even of one of these. Instead we've given a sort of parts list of reactions that, if applied repeatedly and in the appropriate order to any of the possible oligosaccharides would completely break it down, so we do catalog all the gene products and molecular functions needed for catabolism but without really describing the whole process. @ukemi Do we declare victory at this point or defer for future discussion?
• Check the beta-galactosidase involvement in this pathway. It might be ok, because linker chain 2 would be an appropriate substrate.
  Is this an objection to the final reaction in the Reactome pathway, GLB1 hydrolyses linker chain(2)? Or is it OK, at least for now, in which case we can move deferred items into a separate TABLED issue and close this one? @ukemi
• Is the blurb for R-HSA-2090043 correct? Now it is, and thinking about how to fix it led to -
• RHEA reactions that the Reactome reactions should be conformed to: R-HSA-1678660 : RHEA:15125. Done.
• More RHEA - Reactome conforming. DEFER for now. For other reactions, I can't find RHEA reactions that exactly correspond to Reactome one, so defer additional fixes for now. Accounting for extent of HS (de)polymerization consistently in RHEA and Reactome may be tricky / impossible but we can at least mine out correct released monomers, then as a future project go back to RHEA to discuss handing the (de)polymerization issue. For the distant future: would some kind of narrow / broad synonym idea work to relate Reactome and RHEA instances?
• Make mouse model production