Fig. 2: Prototyping and optimization of modules 1–3.

a, Optimizing the CO₂-fixation efficiency of module 1 (CETCH cycle). Shown as CO₂ molecules fixed per molecule of starting substrate (that is, propionyl-CoA) over 2 h for the original CETCH cycle⁵ and the optimized CETCH cycle (CETCH 5.5, this study). In CETCH 5.5, ATP regeneration was changed from a polyphosphate- to a creatine phosphate-based system. b, Prototyping module 2. BHAC efficiency (that is, amino group recycling) was tested by providing different glyoxylate concentrations at limiting glycine concentrations. Reactions were started with 0, 250, 500, 750 or 1,000 µM glyoxylate, while the glycine concentration was fixed at 250 µM. Oxaloacetate was converted into malate through Mdh. Reactions were stopped after 60 min, and malate formation was quantified by LC–MS. In the absence of any amino group cycling (that is, if imminosuccinate were completely hydrolyzed into oxaloacetate), malate yield would be limited to 250 µM at glyoxylate concentrations >500 µM (dashed line). The assays with 750 and 1,000 µM glyoxylate indicate almost 90% amino group recycling (Extended Data Fig. 1). c, Prototyping and optimization of modules 1–3. Shown are the acetyl-CoA (points) and malonyl-CoA (triangles) yields of combined modules 1–3, when using creatine phosphate-based (purple) or polyphosphate-based (petrol) ATP regeneration. Reactions in a and c were started with 100 µM propionyl-CoA. Experiments with phosphocreatine-based ATP regeneration included 2 U ml⁻¹ creatine phosphokinase. Experiments using polyphosphate-based ATP regeneration used 0.5 U ml⁻¹ polyphosphate kinase. Conversion of acetyl-CoA to malonyl-CoA was catalyzed by addition of 100 mU ml⁻¹ propionyl-CoA carboxylase D407I (Pcc*, 26; Fig. 1d). All experiments were performed in technical triplicates and are displayed as mean ± s.d., except CETCH 5.4 assays (data were obtained from the original publication⁵).

Source data