keratan sulfate biosynthesis (R-HSA-2022854)

[ukemi]

• Create a mouse model based on the Reactome import.
• R-HSA-2046298 is missing a preceding reaction in the import, but it is in the Reactome diagram. PD comment: diagrams are laid out manually and a curator can choose to lay out reactions involving shared inputs and outputs to imply directly-provided-input relatioships or not, as the curator wishes. In this case, it looks like there is a preceding event, R-HSA-2046265, that the curator showed in the diagram but did not annotate - now fixed, so diagram and annotation agree.

Funderburgh JL. Keratan sulfate biosynthesis. IUBMB Life. 2002;54(4):187-194. PMID: 12512857 looks like a good starting point for sorting out all of additional issues listed -

• An initial set of reactions to synthesize a linker is needed for keratan sulfate biosynthesis, parallel to the ones for chondroitin sulfate and heparan sulfate (R-HSA-1971475)- Linker synthesis (R-HSA-1971475) #166 - but from Figure 1 in the Funderburgh paper, it's a different linker, including the mannose residues involved in KS branching (next bullet point)not present in the chondroitin / heparan linker, so a new reaction sequence is needed and PMID:29625490 is no use here as it focuses on the role of a specific BGALT enzyme in the elongation reactions that build on that linker, and I can't find any references in PubMed for the synthesis of the KS linkers. Mannose residues might be added using GDP-mannose as a donor, suggesting that RHEA reactions that consume GDP-mannose might be a lead - no luck there either. Is it time to DEFER annotation of the steps involved in forming KS linkers until someone does the experiments to work them out? And to patch the Reactome blurb / summation to say that what we can annotate is basically the content of Funderburgh figure 3 (and also as the normal process that is disrupted by loss of BGALT7 in Figure 7 of PMID: 29625490, which shows the core sequence of steps that, repeated a large and variable number of times, can generate the diverse KS moieties found on real proteins.
• I’m not certain about the order of events for R-HSA-2046265 and R-HSA-2046298. The diagram in Reactome makes it look like the elongation of an antenna is dependent elongation of the chain itself. But the B4galt would work on a GlcNac, not a Gal. It just looks fishy to me. I think it’s because there is no reaction that results in a branch? It turns out the branch is due to the mannose in the linker region (PMID:29625490).
• Ontology ticket- Change the name of [GO:0046964] 3'-phosphoadenosine 5'-phosphosulfate transmembrane transporter activity to 3'-phosphonato-5'-adenylyl sulfate transmembrane transporter activity? 3'-phosphonato-5'-adenylyl sulfate is the Chebi name for ChEBI:58339. If not changed we should at least add a synonym. @ukemi I'd like to tag Harold on this but GitHub won't let me. Do we need to add him to a list of legal targets for this part of GitHub?
• In the assignment of paralogs and their functions should we use papers like PMID:14706853 to pick only the most likely one? MAYBE: doing it right requires enough expertise in the relevant enzymology and carbohydrate chemistry to figure out the exact correct mappings of specific gene product enzymes to specific substrates or, more likely, families of substrates. And can we extrapolate reliably from specificity in vitro to specificity in vivo? And will a good catalog also cause a combinatorial explosion?
• We need to have a look at specificity of keratan sulfotransferase activity. First, its target in the oligosaccharide. It apparently can work on both the GlcNac and the Gal residues. But there is also GO:0001517 that works on the GlcNac. There is nothing specific for the Gal.
• We need to have a look at specificity of keratan sulfotransferase activity. Second, the role of the protein to which the oligosaccharide is attached (quotes are from UniProt entries):
  CHST2 - "keratan-like structures on N-linked glycans and within mucin-associated glycans"
  CHST5 - "non-reducing N-acetylglucosamine (GlcNAc) residues and O-linked sugars of mucin-type acceptors. ... Has no activity toward keratan."
  CHST6 - "position 6 of non-reducing N-acetylglucosamine (GlcNAc) residues of keratan" But also "Cooperates with B4GALT4 galactosyltransferase and B3GNT7 N-acetylglucosaminyltransferase to construct and elongate the sulfated disaccharide unit [->3Galbeta1->4(6-sulfoGlcNAcbeta)1->] within keratan sulfate polymer. Involved in biosynthesis of keratan sulfate in cornea, with an impact on proteoglycan fibril organization and corneal transparency (PubMed:17690104, PubMed:11278593, PubMed:12218059)." - Is this the major sulotransferase activity for our purposes in this pathway?
• It looks like the pathway we have describes KSI. Should we make pathways for KSII and III? PMID:12512857, PMID:32623943