Response to letter-to-the-editor: “Acetyl-CoA synthetase (ACSS2) does not generate butyryl- and crotonyl-CoA”

Dear Editors of Molecular Metabolism,

Thank you for the opportunity to respond to the Letter-to-the-Editor regarding our recent publication on the substrate specificity of acetyl-CoA synthetase short chain 2 (ACSS2) and to provide clarification on certain aspects of our study.

The authors of this comment suggest that detecting short-chain acyl-CoAs (SCA-CoAs) through the regulation of ACSS2 expression in living cells might offer a more reliable approach to evaluating its role in SCA-CoA production. We respectfully disagree. ACSS2 depletion and the resulting drop in acetyl-CoA concentration would indirectly affect the cellular concentration of several SCA-CoAs, including crotonyl-CoA. Indeed, acetyl-CoA produced by ACSS2 feeds acetyl-CoA carboxylase 1 (ACC1), whose activity is required for histone butyrylation and crotonylation. Furthermore, ACSS2 regulates the expression of genes involved in gluconeogenesis and fat metabolism, thereby potentially impacting SCA-CoA production significantly. Thus, in our opinion, altering ACSS2 levels in cells would not provide conclusive evidence for its substrate specificity, which was the primary focus of our study.

By far the most reliable way to assess the ability of ACSS2 to generate SCA-CoAs from the corresponding short-chain fatty acids (SCFA) is through in vitro assays with purified components. Although it is acknowledged that in vitro conditions, despite closely mimicking in vivo environments as in our study, might impact the specific activity of an enzyme, it is highly improbable that they would affect its substrate specificity, which constituted the main objective of our research. To further minimize any issue with individual acetyl-CoA synthetase preparations, we used human and yeast enzymes from different sources.

The authors of the comment further raise doubts regarding our study's conclusion that ACSS2 cannot generate butyryl- or crotonyl-CoA, speculating that a low concentration of these molecules in our assay may have escaped detection. Importantly, our mass spectrometry covers an SCA-CoA concentration range that is linear for at least 4 orders of magnitude and has a detection limit for SCA-CoAs of <0.5 to <10 nM, depending on the species. Since our assay converted up to half of 0.5 mM acetate into acetyl-CoA, a 1000-fold lower efficiency for crotonyl-CoA as suggested by the authors of the comment would yield 0.5 μM, still largely above the detection limits. Thus, at the analyzed SCFA concentrations of 0.5 mM, we can exclude production of butyryl- and crotonyl-CoA. This concentration was deliberately chosen to be much higher than the intracellular levels of SCFA. The concentration of acetate, by far the most abundant species, is 0.05–0.2 mM in the plasma and is unlikely to be higher within cells, consistent with the reported K_m value of ACSS2 for acetate of about 0.05–0.11 mM. If trace amounts of other SCA-CoAs were to be generated using excessively higher SCFA concentrations, for which we have no data, this result would not be physiologically relevant. Conclusions similar to ours were already reached previously with commercially available or self-purified enzyme, albeit using less sensitive methodology. For instance, Frenkel and Kitchens reported that acetyl-CoA synthetase purified from Baker's yeast had high substrate specificity for acetate and propionate. Patel and Walt found that commercially purchased enzymes accepted 3-chloropropionic acid, similar in size to butyric acid, as a substrate, but not butyric acid itself. This finding provides evidence that these enzymes can maintain their activity on substrates other than acetate, but are highly selective in their choice of longer-chain fatty acid molecules.

Mindful of the potential limitations of individual methodologies, we combined three independent approaches – cellular, in vitro and in silico – which consistently support the conclusions of our study. We feel that it is crucial to convey this information to the community. We suggest that the indirect effects of ACSS2 knockdown should be reevaluated based on our findings and those reported in the existing literature, rather than accepting the hypothesis of its broad substrate specificity as an established fact.