The claim
Santana-Molina and colleagues (THEORY-1109, as read in NOTE-1239) find Asgard-archaeal enzymes mainly in cytosolic glycolysis and the pentose phosphate pathway, and alphaproteobacterial ones mainly in the mitochondrial TCA cycle and pyruvate conversion. Any model in which an archaeal host that kept its own glycolysis took up a bacterium that brought a TCA cycle predicts that split, whether the partnership began in syntrophy or by engulfment by an already complex host. What the split does count against is a host whose own carbon metabolism was replaced by the symbiont’s, the view the authors set themselves against, citing Martin and Müller (NOTE-1239); that view is itself a syntrophy model in the record’s grouping (NOTE-1238), so the split tells syntrophic models apart without singling syntrophy out. The authors themselves list what would discriminate the models, “the types of metabolites exchanged, … the timing of mitochondrial acquisition, the mechanism of mitochondrial uptake”, and test none of them.
What it does not say
Not that syntrophy is wrong. The claim is that this evidence does not tell it from its rivals; the syntrophic thesis is THEORY-1109’s wider form, Proposed.
Note of 2026-10-30: the clause narrowed out
The title said “any origin in which an archaeal host took up a bacterium with a TCA cycle predicts it”. By NOTE-1239 and THEORY-1109, the split counts against an origin of that kind, one in which the host’s carbon metabolism was replaced by the symbiont’s, so the clause was too wide (ADR-047 kind 3). The title now names the hosts it holds for.