looking for evidences for MAR03458 and 3Z,7Z,10Z-Hexadecatrienoyl Coenzyme A

[ANiknejad]

Hello Human-GEM Team (maybe @Devlin-Moyer )

I am looking for 3Z,7Z,10Z-Hexadecatrienoyl Coenzyme A , that should be a substrate for ECI1/2 and DECR1/2 (weird, knowing that DECR1 is mitochondrial, while DECR2 is peroxisomal)

see HPA https://v19.proteinatlas.org/ENSG00000177646-ACAD9/tissue/metabolic/Omega-6+fatty+acid+metabolism

I do not find any evidence for 3Z,7Z,10Z-Hexadecatrienoyl Coenzyme A existing, could you confirm that reactions MAR03458 and MAR03457 should exist?

Anne

Anne Niknejad
Senior biocurator
SIB | Swiss Institute of Bioinformatics Unil, Department of Ecology and Evolution - Quartier Sorge, bâtiment Biophore - 1015 Lausanne t: +41 21 692 42 21 - f: +41 21 692 41 65 Anne.Niknejad@sib.swiss - https://www.sib.swiss

[Devlin-Moyer]

This paper claims that DECR1 (which they refer to as "mDECR") "reduces a wide range of substrates of varying lengths of aliphatic tail and with different stereochemical conformations" and that "[t]he active site pocket is large enough to accommodate substrates with chain lengths exceeding [] six carbons". That is basically the only paper I can find that comments on the substrate chain length preferences of human DECR1, so while it's not direct evidence that DECR1 can reduce 2E,4Z,7Z,10Z-hexadecatetraenoyl-CoA to 3Z,7Z,10Z-hexadecatrienoyl-CoA, I also can't find any evidence that 2E,4Z,7Z,10Z-hexadecatetraenoyl-CoA is too long/large of a substrate for DECR1. As outlined in #761, 2E,4Z,7Z,10Z-hexadecatetraenoyl-CoA is one of the expected intermediates of the beta-oxidation of gamma-linolenic acid, which definitely seems to be present in human cells (1, 2, 3).

With regard to compartmentalization, #788 ensured that the GPR of MAR03457 only includes DECR1 and not also DECR2 (it also ensured that the GPR of MAR03471, the peroxisomal equivalent of MAR03457, only includes DECR2 and not DECR1).

[ANiknejad]

Thanks a lot, @Devlin-Moyer, for these clear explanations!

[ANiknejad]

Hello @Devlin-Moyer

Still some questions, actually:

in ticket #786 you annotated 3E for the beta-oxidation of 8Z,11Z-20:2, so how can we assess that DECR1 can produce 3Z,7Z,10Z-Hexadecatrienoyl Coenzyme A and not 3E,7Z,10Z-Hexadecatrienoyl Coenzyme A ?

We will then have beta-oxidation of 6Z,9Z,12Z-18:3 (gamma-linolenic) by DECR1 to produce 3Z while beta-oxidation of eicosadienoate by DECR1 to produce 3E?

Could you please explain how the enzyme would decide the stereo ?

Many thanks in advance for your patience,

Anne

[Devlin-Moyer]

Strictly speaking, DECR1 should reduce 2E,4Z,7Z,10Z-Hexadecatetraenoyl-CoA to 3E,7Z,10Z-Hexadecatrienoyl-CoA (see the introduction and figure one of the same paper that I cited in my previous comment). However, the ECIs can reversibly isomerize between 3Z-enoyl-CoAs, 2E-enoyl-CoAs, and 3E-enoyl-CoAs (source), so the 3E,7Z,10Z-Hexadecatrienoyl-CoA that is directly produced by DECR1 probably exists in equilibrium with 2E,7Z,10Z-Hexadecatrienoyl-CoA as well as 3Z,7Z,10Z-Hexadecatrienoyl-CoA. It would probably be ideal to have MAR03457 produce 3E,7Z,10Z-Hexadecatrienoyl-CoA instead of 3Z,7Z,10Z-Hexadecatrienoyl-CoA, but changing that wouldn't really change anything about the predictions that the model makes, and Human-GEM already has tons of other reactions that merge several distinct real-life reactions into one Human-GEM reaction. I remember coming across several other cases like this and deciding to focus on the larger errors/inconsistencies and just not mentioning the 3Z/3E inconsistencies in the various issues that I've opened about fatty acid metabolism