search

SysBioChalmers / Human-GEM

The generic genome-scale metabolic model of Homo sapiens
https://sysbiochalmers.github.io/Human-GEM-guide/
Creative Commons Attribution 4.0 International
97 stars 41 forks source link

SLC15A19 Is Not A Nucleotide Transporter #695

Open Devlin-Moyer opened 1 year ago

Devlin-Moyer commented 1 year ago

While working on #694, I noticed a large number of transport reactions that transport (d)NTPs and (d)NDPs between the cytosol and mitochondria that are associated with SLC25A19 (ENSG00000125454). There is a warning at the top of the Uniprot page for SLC25A19 that says that, while it was initially claimed to be a mitochondrial nucleotide transporter, it was eventually shown to only be capable of transporting thiamine diphosphate.

All but 2 of the 56 reactions currently associated with SLC25A19 transport (d)NTPs and (d)NDPs between the cytosol and mitochondria, and I believe they should all be removed, since other existing reactions and the new ones proposed in #694 will ensure that all (d)NTPs and (d)NDPs can still move between the [c] and [m] compartments.

The two reactions that don’t involve nucleotides are MAR04933: AKG [e] + Na+ [e] ⇒ AKG [c] + Na+ [c] and MAR05047: Pi [l] ⇒ Pi [c]. MAR04933 is nearly identical to MAR05992, which uses 3 sodium ions instead of 1 and is associated with SCL13A3, which is actually known to catalyze that reaction according to Uniprot, so I think MAR04933 should be removed. MAR05047 appears to be the only way for inorganic phosphate to move from the lysosome to the cytosol, so I think it should be kept but have SLC25A19 removed from its GPR. As far as I can tell, people have known that phosphate is exported from human lysosomes since at least 1991 (source), but never determined which protein is responsible — this paper from 2021 cited that 1991 paper and no others when it mentioned lysosomal phosphate transport.

Proposed Changes

haowang-bioinfo commented 1 year ago

could it be good enough to turn the GPRs of these reactions into blank?

Devlin-Moyer commented 1 year ago

Maybe, but I feel like that would be inconsistent with how we resolved issues like #568, #569, #570, and #679. Those all identified reactions that transported the same metabolite between the same compartments that misrepresented how certain transporters worked, and we either removed or edited the reactions that did not correspond to the activity of a known transporter, rather than just removing the GPRs but keeping those reactions. Also, as I mentioned in the initial comment, the new reactions proposed in #694 would allow transport of all of the metabolites involved in all of these reactions to move between the same compartments, and we have evidence to support that those particular reactions are actually mediated by a specific transporter, while we have no evidence that any of the reactions I mentioned here actually reflect the activities of any known transport proteins.