Auto-converted DRAFT module for Isoleucine biosynthesis, mined from the GapMind amino-acid pathway definition 'ile.steps' and grounded against ModelSEED. Generated by a feasibility-spike importer; GO molecular-function term assignments and biological prose are intentionally left for human/deep-research review and must not be treated as curated.
MINED DRAFT -- not curated. Source: GapMind .steps (PaperBLAST, Price/Arkin LBL). Steps map to annotons; GapMind OR-rules map to EXACTLY_ONE variant_sets; EC numbers grounded to ModelSEED/KEGG/MetaCyc. 'ignore'/'term' GapMind tokens were dropped.
All recommended fields populated.
✗ none found
No MODULE:gapmind_ile_biosynthesis deep-research report alongside the module YAML.
10 leaf node(s) with no concrete protein grounding:
✓ every declared conforms_to bundle matches its template motif.
No concrete UniProt-grounded genes in this module.
MetaCyc L-isoleucine biosynthesis II describes the formation of 2-oxobutanoate via citramalate. The other steps are the same (although it gives a different EC number for ilvC because of different cofactor preference) The citramalate synthase from Leptopsira interrogans (LA_2350, NP_712531, or Q8F3Q1) has been characterized biochemically but is not in the curated databases, see PMID:18498255 The putative citramalate synthase HVO_0644 (D4GSQ2) from Haloferax volcanii is required for isoleucine biosynthesis, see PMC4300041.
In leucine synthesis, LeuCD allows the dehydration of 2-isopropylmalate and hydration to 3-isopropylmalate. Similarly, many of these enzymes allow the isomerization of citramalate to 3-methylmalate via citraconate. Citramalate isomerase is usually given as EC 4.2.1.35, as opposed to 4.2.1.33 for traditional leuCD. However, many the bacteria with the citramalate pathway appeared to have "typical" leuBCD (i.e., Desulfovibrio vulgaris Hildenborough, Desulfovibrio vulgaris Miyazaki F, Bacteroides thetaiotaomicron, Magnetospirillum magneticum AMB-1, and Synechococcus elongatus PCC 7942). So we do not try to distinguish between 3-isoprpylmalate dehydratase and citramalate isomerase. Ignore a 2,3-methylmalate dehydratase (Q0QLE2,Q0QLE1) which is >50% identical to leuCD from DvH (DVU2982,DVU2983). Ignore some BRENDA annotations without subunit information, and ignore CharProtDB::CH_122621 (leuCD fusion) which is not actually characterized. DvH leuC (DVU2982) has similarity to both LeuC and to homoaconitase, and fitness data confirms its role in amino acid biosynthesis, so explicitly include it. CA265_RS15830 (uniprot:A0A1X9Z7T5) from Pedobacter sp. GW460-11-11-14-LB5 is important for fitness unless amino acids are added. uniprot:S3E7P8 is annotated as this in SwissProt but we did not find experimental evidence, so it is ignored.
A mutant of BAC65258.1 (uniprot:Q845W4) was shown to be a leucine auxotroph in PMC155387. CA265_RS15840 (uniprot:A0A1X9Z766) from Pedobacter sp. GW460-11-11-14-LB5 is important for fitness unless amino acids are added.
The short protein uniprot:P87267 is misannotated as this in BRENDA, so it is ignored.
uniprot:Q8ZKF6 is very similar to E. coli acs and likely has this activity, so it is ignored.
alpha-ketobutyrate synthase or 2-oxobutanoate:ferredoxin oxidoreductase (in reverse) is a heterodimeric enzyme
[imported from val.steps]
The isolated catalytic subunit can have some activity on its own, so it's not clear if the regulatory subunit (ilvH) is always required, but ilvH does always seem to be present. uniprot:P0ADG1 is annotated with this EC number but not explicitly as the small regulatory subunit, so it was added manually. uniprot:Q93YZ7 is annotated as this but without the EC number, so is added manually. Most regulatory subunits have an N-terminal ACT domain and a C-terminal ACT-like domain, but E. coli IlvM, which is required for the activity of E. coli acetohydroxyacid synthase isoenzyme II, has the N-terminal ACT domain only. We identified several other short (one-domain) regulatory subunits. In Rhodanobacter and related genera, the putative regulatory subunit has just one ACT domain (i.e., LRK54_RS10305, which is nearly identical to A0A154R0Y7). Based on sequence analysis, short ilvH probably maintains the ability to bind valine and to bind the catalytic subunit, but not the ability to bind ATP or other regulatory subunits. Mutant fitness data confirms that LRK54_RS10305 is involved in amino acid biosynthesis. In Xanthomonas campestris, the one ACT-domain protein Xcc-8004.1058.1 (uniprot:A0A0H2X4P1), which is conserved next to ilvI, has a similar fitness pattern as ilvI (Alice Castaing, unpublished data). Furthermore, its AlphaFold structure is very similar to that of E. coli IlvM (TM-score 0.89, RMSD 1.49 A, foldseek). So Xcc-8004.1058.1 is another short regulatory subunit. Similarly, in Brevundimonas sp. GW460-12-10-14-LB2, the putative ilvH has the ACT domain only (Brev2_1981 = A0A161J739). Many Thermoproteota seem to have a diverged short regulatory subunit, such as the ACT domain protein KCR_RS03285 (uniprot:A0A7J3AYJ4), which is conserved next to ilvI. Foldseek shows that this protein is similar to the ACT domain of (p)ppGpp synthase but also to IlvH of Staphylococcus aureus, so we predict that it is the regulatory subunit. Many Chryseobacterium have a truncated 1-domain regulatory subunit (i.e., uniprot:A0A316WEF5) that is similar to the N-terminal part of a regulatory subunit from Pedobacter (CA265_RS15810, which was confirmed by fitness data); because some of these strains of Chryseobacterium grow in the absence of any one of the branched-chain amino acids (i.e., GW821-FHT04C04 and GW821-FHT04A06), the 1-domain protein is presumably functional. [imported from val.steps]
The three EC numbers correspond to different preferences for NAD(P)H as the cofactor; the transformations to the carbon skeleton are the same. CH_123630 is added because it is annotated as this but with no EC number. [imported from val.steps]
IlvD is involved in the biosynthesis of isoleucine, with (R)-2,3-dihydroxy-3-methylpentanoate as the substrate, and in the biosynthesis of valine and leucine, with (R)-2,3-dihydroxy-3-methylbutanoate as the substrate. The ignored enzyme is involved in salinosporamide A biosynthesis but does a very similar reaction and is >50% identical to N515DRAFT_0569, which is confirmed by fitness data to be biosynthetic [imported from val.steps]
Q8NS92 is ignored because it is primarily a transcriptional regulator. Similarity to aromatic amino acid transaminases or tyrosine transaminases is ignored as they often are often non-specific.