Problems with Beta-Oxidation of Tocopherols and Tocotrienols #738

While looking for mtiochondrial reactions that were associated with peroxisomal genes in #634, I noticed some issues with the beta-oxidation pathways for alpha- and gamma-tocopherol and tocotrienol:

For reference, here are the structures of alpha- and gamma-tocopherol and tocotrienol:

Those GPRs are all mixtures of enzymes that participate in mitochondrial and peroxisomal β-oxidation, but all of these reactions involve exclusively mitochondrial metabolites.There are other problems with these reactions, some of which I address in #791, and the rest of which I think will be easier to address once we decide if we're moving some of them to the peroxisomal compartment or keeping everything in the mitochondrial compartment.

Expected behavior:

It seems like people aren't sure whether the first two rounds of beta-oxidation of tocotrienols and tocopherols (i.e. MAR06432, MAR06433, MAR06482, MAR06484, MAR06447, MAR06448, MAR06470, and MAR06471) occur in the mitochondria or peroxisomes of human cells; see the text under the subheading "Intracellular compartmentation of vitamin E metabolism" and Figure 3 of this paper. If we want to keep them in [m], they should be associated with ACADL and HADHA and HADHB, since ACADL appears to be the only mitochondrial acyl-CoA dehydrogenase that can act on substrates that are both branched and unsaturated (sources: 1, 2), and the HADHA/HADHB complex can perform all remaining rounds of beta-oxidation on acyl-CoAs with at least 10 carbons in their primary chains (source). If we want to move them to [x], they should instead be associated with (ACOX2 or ACOX3) and (EHHADH or HSD17B4) and SCP2 (sources: 1, 2), and we'd also need to add in transport reactions to move the 9' intermediates from [x] to [c] and then [c] to [m].

The next two rounds of beta-oxidation (from the 9' intermediates to the 7' intermediates) that people are apparently confident only happen in mitochondria should also be associated with ACADL, since it appears to be capable of recognizing acyl-CoAs with primary chains as short as 6 carbons long (figure 4G from this paper) and continues to be the only mitochondrial ACAD that recognizes acyl-CoAs that are both unsaturated and branched. The final round of beta-oxidation (i.e. MAR06436 and MAR06453) should instead be associated with ACADSB, since, at that point, all of the double bonds from the tocotrienols have been removed, and the side chains are less than 6 carbons long sources: table 2 from this paper and figure 4H from this paper)

Since the 9', 7', and 5' intermediates are all oxidized in mitochondria and have side chains that are less than 10 carbons long, their beta-oxidation reactions should be associated with ECHS1, HADH, HSD17B10, and SCP2 rather than HADHA and HADHB (source). Note that I'm intentionally excluding ACAA1 and ACAA2 from these GPRs -- apparently, they only recognize straight-chain acyl-CoAs (source), and all of these intermediates are branched.

ACADL, ACOX2, and ACOX3 can oxidize (2S)-methylacyl-CoAs but not (2R)-methylacyl-CoAs (sources: 1, 2), and naturally-occurring tocopherols are always the all-R stereoisomers. AMACR can isomerize (2R)-methylacyl-CoAs to (2S)-methylacyl-CoAs, but doesn't touch methyl groups at any other position, so it should only be associated with the oxidation reactions for the 9'- and 5'-chromanol intermediates, i.e. MAR06486, MAR06472, MAR06436, and MAR06453. The 9' tocotrienol intermediates don't have stereocenters where the methyl groups are attached because of their double bonds.

Each of these reactions seems to combine a full round of beta-oxidation with a CoA activation step and a CoA hydrolysis step. I found it difficult to figure out which acyl-CoA synthetases and thioesterases could recognize branched and unsaturated substrates, so my best suggestion at the moment is to associate ACSL1 or ACSL5 with the first round of beta-oxidation (just ACSL1 if it's moved to the peroxisome), ACSM3 or ACSM4 with the second round, ACSM1 or ACSM3 or ACSM4 for the third and fourth rounds, and just ACSM1 for the final round (source). THEM4 seems like it could hydrolyze any of these CoA intermediates in mitochondria (source), and ACOT8 could probably handle the 11' and 9' intermediates in peroxisomes (source)

Proposed changes:

If moving the first two rounds of beta-oxidation to the peroxisomal compartment:

• Replace all metabolites in MAR06432, MAR06433, MAR06482, MAR06484, MAR06447, MAR06448, MAR06470, and MAR06471 with their peroxisomal equivalents
• Change the GPRs of MAR06432, MAR06433, MAR06482, MAR06484, MAR06447, MAR06448, MAR06470, and MAR06471 to ACSL1 and (ACOX1 or ACOX3) and (EHHADH or HSD17B4) and SCP2 and ACOT8
• Create new transport reactions to move 9-carboxy-alpha-tocotrienol, 9-carboxy-alpha-chromanol, 9-carboxy-gamma-tocotrienol, and 9-carboxy-gamma-chromanol from [x] to [c] and [c] to [m]

If keeping all reactions in the mitochondrial compartment:

• Change the GPRs of MAR06432, MAR06482, MAR06447, and MAR06470 to (ACSL1 or ACSL5) and ACADL and HADHA and HADHB and THEM4
• Change the GPRs of MAR06433, MAR06484, MAR06448, and MAR06471 to (ACSM3 or ACSM4) and ACADL and HADHA and HADHB and THEM4

Either way:

• Change the GPRs of MAR06434, MAR06486, MAR06450, and MAR06472 to (ACSM1 or ACSM3 or ACSM4) and AMACR and ACADL and ECHS1 and (HADH or HSD17B10) and SCP2 and THEM4
• Change the GPRs of MAR06435, MAR06488, MAR06451, and MAR06473 to (ACSM1 or ACSM3 or ACSM4) and ACADL and ECHS1 and (HADH or HSD17B10) and SCP2 and THEM4
• Change the GPRs of MAR06436 and MAR06453 to ACSM1 and AMACR and ACADSB and ECHS1 and (HADH or HSD17B10) and SCP2 and THEM4