3-hydroxypropionate bicycle

The 3-hydroxypropionate bicycle of Chloroflexus aurantiacus, which fixes two bicarbonate molecules per turn using two biotin-dependent carboxylases and no oxygen-sensitive ferredoxin chemistry at all. The first cycle carboxylates acetyl-CoA to malonyl-CoA, reduces that to 3-hydroxypropionate, reactivates and reduces it to propionyl-CoA, carboxylates again to methylmalonyl-CoA, and rearranges that to succinyl-CoA, which is converted through to (S)-malyl-CoA and cleaved to release glyoxylate and regenerate acetyl-CoA. The second cycle exists because glyoxylate is not directly usable: it is condensed with propionyl-CoA and processed to citramalyl-CoA, whose cleavage yields pyruvate for biosynthesis and returns acetyl-CoA to the first cycle. Hence "bicycle" - two cycles sharing the propionyl-CoA node. Being oxygen-insensitive, it works in the aerobic and microaerophilic conditions that exclude MODULE:reductive_tca_cycle and MODULE:wood_ljungdahl_pathway, with which it is a sibling, as it is of MODULE:calvin_benson_bassham_cycle. The distinct archaeal 3-hydroxypropionate/4-hydroxybutyrate cycle of the Sulfolobales is a different pathway and is not modelled here.

Grounded on Chloroflexus aurantiacus. Two honest gaps are recorded rather than papered over. First, there is no GO molecular-function term for malonyl-CoA reductase (EC 1.2.1.75): the closest candidates are the malonate-semialdehyde dehydrogenase terms, which describe only half of this bifunctional enzyme's chemistry, so the annoton carries a free-text function and no `term`, and the ontology gap is the one thing in this module worth raising with GO. Second, propionyl-CoA synthase (Pcs) could not be resolved to a UniProt accession, and the search below was run twice rather than left at a first attempt. Entries named "Propionyl-CoA synthase" do exist, but all are 218-aa environmental Chloroflexus sp. records flagged `Fragment` at protein-existence level 4, carrying only PF00501 (AMP-dependent synthetase/ligase) - that is a fragment of the first of the enzyme's three activities, not the trifunctional protein, and none is C. aurantiacus. Searching EC 6.2.1.36 returns only archaeal 3-hydroxypropionyl-CoA synthetases from the unrelated 3-HP/4-HB cycle. C. aurantiacus does contain a 1822-aa protein (UniProtKB:A9WEU4) annotated "acetate--CoA ligase", a length consistent with the trifunctional architecture and a first activity consistent with the name, but that is a hypothesis and is deliberately NOT asserted as the representative member: nothing ties it to the characterised Pcs beyond length and domain class. Recorded here so the next attempt starts from this rather than repeating it. Pcs is nonetheless modelled, because its three activities each have GO terms (GO:0043955, GO:0043956, GO:0043957) and the trifunctional architecture is well established. The shared steps between succinyl-CoA and (S)-malyl-CoA, and the mesaconyl-CoA isomerization and hydration of the second cycle, are real but not separately decomposed: they are conventional CoA-ester chemistry and not diagnostic of the pathway. Several second-cycle enzymes are bifunctional or trifunctional in this organism, so counting proteins will not give the number of steps.