Design and implementation of aerobic and ambient CO2-reduction as an entry-point for enhanced carbon fixation

Satanowski, Ari; Marchal, Daniel G.; Perret, Alain; Petit, Jean-Louis; Bouzon, Madeleine; Döring, Volker; Dubois, Ivan; He, Hai; Smith, Edward N.; Pellouin, Virginie; Petri, Henrik M.; Rainaldi, Vittorio; Nattermann, Maren; Burgener, Simon; Paczia, Nicole; Zarzycki, Jan; Heinemann, Matthias; Bar-Even, Arren; Erb, Tobias J.

doi:10.1038/s41467-025-57549-4

Download PDF

Article
Open access
Published: 01 April 2025

Design and implementation of aerobic and ambient CO₂-reduction as an entry-point for enhanced carbon fixation

Nature Communications volume 16, Article number: 3134 (2025) Cite this article

8066 Accesses
4 Citations
28 Altmetric
Metrics details

Subjects

Abstract

The direct reduction of CO₂ into one-carbon molecules is key to highly efficient biological CO₂-fixation. However, this strategy is currently restricted to anaerobic organisms and low redox potentials. In this study, we introduce the CORE cycle, a synthetic metabolic pathway that converts CO₂ to formate at aerobic conditions and ambient CO₂ levels, using only NADPH as a reductant. Combining theoretical pathway design and analysis, enzyme bioprospecting and high-throughput screening, modular assembly and adaptive laboratory evolution, we realize the CORE cycle in vivo and demonstrate that the cycle supports growth of E. coli by supplementing C1-metabolism and serine biosynthesis from CO₂. We further analyze the theoretical potential of the CORE cycle as a new entry-point for carbon in photorespiration and autotrophy. Overall, our work expands the solution space for biological carbon reduction, offering a promising approach to enhance CO₂ fixation processes such as photosynthesis, and opening avenues for synthetic autotrophy.

Construction and modular implementation of the THETA cycle for synthetic CO₂ fixation

Article Open access 20 December 2023

Light-driven biosynthesis of volatile, unstable and photosensitive chemicals from CO₂

Article 25 May 2023

The potential of CO₂-based production cycles in biotechnology to fight the climate crisis

Article Open access 01 November 2023

Introduction

The biological conversion of inorganic carbon into biomass is at the core of the global carbon cycle, and key to establish a sustainable (bio)economy, in which the greenhouse gas carbon dioxide (CO₂) serves as the main carbon source. In chemical terms, formation of biomass from CO₂ requires the reduction and condensation of the one-carbon (C1) molecule CO₂ into multi-carbon compounds. This can in principle happen via two different routes: Most CO₂-fixing organisms use “carbon-fixation” routes, in which CO₂ is first incorporated into an acceptor molecule, subsequently reduced and directly converted into a multi-carbon compound through the same pathway¹. Alternatively, CO₂ can be assimilated through a “reduction-first” strategy², in which CO₂ is initially converted into a reduced C1-compound (e.g. formate or carbon monoxide), which is condensed into multi-carbon compounds in a subsequent phase through downstream reactions.

The reduction-first strategy offers several advantages in terms of energetic efficiency and rate, which makes it highly attractive from a biotechnological point of view². Two of the eight naturally existing CO₂-assimilation pathways follow this strategy, namely the reductive acetyl-CoA pathway (also known as the Wood-Ljungdahl pathway) and the reductive glycine pathway^1,3,4,5,6. However, both routes were exclusively found in organisms that grow anaerobically, in the presence of low-redox potential electron donors (e.g., molecular hydrogen (H₂)), and preferably at high CO₂ partial pressures^5,7. While microbial strains using these strategies have been developed for industry and can be grown under these special conditions⁸, many bioproduction hosts, and especially photosynthetic organisms including crops, are restricted to aerobic environments and atmospheric CO₂ levels. This has raised the question of whether reduction-first routes could, in principle, also be realized under aerobic conditions.

A critical challenge to establish an oxygen-tolerant reduction-first CO₂-fixation is the initial step, namely reduction of CO₂ to formate. This reaction requires a redox potential of roughly −460 mV at ambient CO₂-concentrations (400 ppm CO₂ and physiological pH of 7 (eQuilibrator⁹); E₀′ = −432 mV with 1 bar CO₂¹⁰). Aerobic organisms that do not possess low-redox potential electron donors have to rely on the NAD(P)H/NAD(P)⁺ pair (−320 mV¹⁰). This constrains the direct reduction of CO₂, with an estimated Gibbs energy change (∆G) of +25 kJ/mol (400 ppm CO₂, pH 7 and 1 mM for all other reactants⁹), which is almost impossible to achieve with physiological NADPH/NADP⁺ ratios^{11,12,13,14,15,16,17,18} and additionally comes at very high enzyme cost to reach sufficient flux^18,19,20. This general limitation is supported by a recent study, in which aerobic CO₂-reduction via an NADH-dependent formate dehydrogenase was successfully implemented in vivo, but only allowed for slow growth at strongly elevated CO₂ concentrations >10% (i.e., 0.1 bar, 250× atmospheric level) and required an evolved mutation in the respiratory chain, likely increasing the cellular NADH/NAD⁺ ratio²¹.

In this work, we develop a metabolic architecture that is able to convert CO₂ into formate at fully aerobic conditions, and ambient CO₂-concentrations. To overcome the thermodynamic and kinetic bottlenecks associated with CO₂-reduction, we design several pathways (“mini-cycles”) that are energized by ATP-hydrolysis (Fig. 1A, B) and split up the task of CO₂-reduction into multiple enzymatic steps with favorable thermodynamic driving forces. We use high-throughput in vitro bioprospecting to identify suitable enzyme candidates for our designs and realize one of these cycles, the CORE cycle, which we successfully assemble in a modular fashion in Escherichia coli. Using adaptive laboratory evolution (ALE) allows us to achieve a fully functional CORE cycle that supports growth of E. coli by supplying the organism’s complete C1-metabolism and serine biosynthesis from CO₂. The CORE cycle provides a flexible design that holds the potential to improve the efficiency of natural CO₂-fixation pathways, such as photosynthesis (e.g., by augmenting photorespiration²², Fig. 1D), and—at the same time—also opens the door for establishing synthetic reduction-first autotrophy in the future (Fig. 1C).

**Fig. 1: ATP-driven CO₂-reduction as a novel entry-point for carbon fixation.**

Results

Design and evaluation of O₂-tolerant, NAD(P)H-dependent CO₂-reduction pathways

In a theoretical effort, we first designed ten different hypothetical pathways for the ATP-driven, NAD(P)H-dependent conversion of CO₂ to formate (Supplementary Figs. 1–10, Supplementary Table 1). These pathways include at least four reactions (Fig. 1B): (i) carboxylation (incorporation of CO₂ or bicarbonate), (ii) ATP-hydrolysis, (iii) reduction, and (iv) release of formate. All pathways follow this logic but differ in the number and order of the reactions, the amount of ATP hydrolyzed, the reductant (NADH or NADPH), and the “backbone molecule” (i.e. the acceptor substrate) for the initial CO₂ incorporation (Supplementary Table 1).

We evaluated our ten pathway designs based on multiple criteria (Supplementary Table 1). Specifically, we compared their (i) resource requirements (ATP and reducing equivalents), (ii) pathway length (i.e. number of enzymes), (iii) need for complex enzyme co-factors (e.g. adenosylcobalamin (coenzyme B12)) potentially limiting the implementation in organisms such as E. coli or plants^23,24, and (iv) number of new-to-nature enzyme reactions^25,26,27 that would need to be established to realize the cycles. This analysis identified four promising designs that we considered further for implementation. These pathways all required only one new-to-nature reaction as a key step, after which we initially named them: the aspartate-formate-lyase cycle, the reductive formamide pathway, the oxaloacetate hydrolase cycle, and the β-keto-acid cleavage cycle (detailed in Supplementary Note 1).

In vitro prototyping of key enzyme activities identifies the CORE cycle

We next focused on establishing the missing new-to-nature reaction for each of the four different cycles. To that end, we performed a literature search to identify and test different enzyme candidates that would be able to catalyze the required novel reaction. While we failed to establish activity for three of the reactions in vitro (aspartate formate lyase, carbamoyl-phosphate reductase, oxaloacetate hydrolase), we could successfully demonstrate the β-keto acid cleavage reaction (see below). Thus, we sought to realize the full β-keto acid cleavage cycle, which we will refer to as the CORE cycle (CO₂-reduction cycle; Fig. 1E) in the following.

The CORE cycle achieves a net conversion of CO₂ (bicarbonate) to formate at the expense of 1 ATP and reducing power in the form of 1 NADPH. The pathway can also be envisioned in two other “design variants” which differ in their ATP-cost to activate acetoacetate to acetoacetyl-CoA (0–2 ATP, reaction #4 in Fig. 1E; Supplementary Table 1, Supplementary Fig. 1). The CORE cycle first incorporates bicarbonate via carboxylation of acetyl-CoA into malonyl-CoA, followed by reduction of malonyl-CoA to malonate semialdehyde (MSA) through malonyl-CoA reductase (MCR). Next, a β-keto acid cleavage enzyme (BKACE)^28,29 converts the non-native substrate MSA with acetyl-CoA into acetoacetate and formyl-CoA (Fig. 1F). Conversion of formyl-CoA to the final product formate can occur either via hydrolysis (by thioesterases or non-enzymatically^30,31,32), or in a more energy-conserving manner by transferring the CoA-moiety from formyl-CoA onto acetoacetate via a CoA-transferase. To close the cycle, acetoacetyl-CoA is cleaved by a thiolase, thus regenerating two molecules of acetyl-CoA, one of which enters another round of the cycle, while the other is required for the BKACE reaction.

Identification and in-depth characterization of enzyme candidates for the CORE cycle

All reactions of the CORE cycle apart from the proposed BKACE reaction were already known from other natural pathways. To identify suitable enzymes for these reactions, we tested multiple candidates in vitro (Supplementary Table 2). For the carboxylation of acetyl-CoA (reaction #1, Fig. 1E), we identified an engineered propionyl-CoA carboxylase (Pcc_Me_D407I)³³ that is more stable in vitro and less complex compared to many natural acetyl-CoA carboxylases³³. For the NADPH-dependent reduction of malonyl-CoA to MSA (reaction #2), we considered two options: Archaeal MCRs, and bacterial bi-functional MCRs that catalyze the additional reduction of MSA to 3-hydroxypropionate^34,35,36. We tested candidates from both clades (Supplementary Fig. 11A) and decided to use a truncated version of the bi-functional MCR from Chloroflexus aurantiacus, that comprises only its C-terminal domain catalyzing the reduction of malonyl-CoA to MSA³⁷. For the activation of acetoacetate (reaction #4), we tested four candidates that work either by CoA ligation or CoA transfer. For the β-ketothiolase (reaction #5, acetyl-CoA acetyltransferase), we selected AtoB from E. coli (Supplementary Table 2).

To identify candidates that catalyze the new-to-nature BKACE reaction, we leveraged a previously described collection of BKACE proteins²⁹. We tested 124 enzymes for the ability to convert MSA and acetyl-CoA into formyl-CoA and acetoacetate (Fig. 1F) and identified more than 20 homologs with the desired activity. We screened these candidates with an endpoint assay (Supplementary Fig. 11B and Supplementary Table 3) and characterized the eight most promising candidates in detail using an assay in which we measured acetyl-CoA consumption, formyl-CoA production and acetoacetate production under substrate-saturated conditions (Fig. 2A, Supplementary Fig. 12).

**Fig. 2: Biochemical characterization of BKACE.**

In these assays, BKACE15 from Paracoccus denitrificans performed best (Supplementary Table 4). The physiological role of this enzyme is unknown. However, the genetic context suggests that BKACE15 is involved in the degradation of aromatic compounds, while experimentally confirmed substrates include β-ketohexanoate, 3,5-dioxohexanoate, 5-hydroxy-β-ketohexanoate and β-ketopentanoate²⁹. With acetyl-CoA and MSA, BKACE15 showed a specific activity of 770 ± 30 nmol min⁻¹ mg⁻¹ and strong substrate inhibition above 10 mM MSA. Limiting the analysis to physiologically relevant concentrations in the lower mM range, we determined an apparent K_m value for acetyl-CoA at 0.11 ± 0.01 mM, and MSA at 0.83 ± 0.09 mM (Figs. 2B, C, Supplementary Table 5). However, we also observed that the enzyme was inhibited by its product acetoacetate with a half-maximal inhibitory concentration (IC₅₀) of 1.45 ± 0.16 mM (Fig. 2D, Supplementary Fig. 13). We noted that the IC₅₀ for acetoacetate was in the same range as the K_m value for MSA, thus posing a potential bottleneck, if acetoacetate was not removed quickly downstream of the BKACE reaction.

Structure-function studies explain the reaction mechanism of BKACE

To understand the catalytic behavior of BKACE15, we obtained structures of the apo-form at 1.45 Å (PDB 8RIO), the enzyme in complex with malonate and free CoA at 1.81 Å (PDB 8RIP), and in complex with acetoacetate and acetyl-CoA at 2.10 Å (PDB 9HNF) (Supplementary Fig. 14). BKACE15 forms a homotetramer where each active site is accommodated in a TIM-barrel fold, and consists of eight β-sheets that are surrounded by eight α-helices (Fig. 2E). The two openings of the TIM barrel are closed off by smaller α-helices. As previously reported for the BKACE from Candidatus Cloacamonas acidaminovorans, a close homolog of BKACE15, each active site coordinates a single Zn²⁺ ion²⁸. We confirmed a 1:1 zinc binding ratio in BKACE15 using Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES, Supplementary Fig. 15). Each active site also exhibits only one entry channel for both substrates that appears fully occupied upon acetyl-CoA binding (Fig. 2F, G). This suggests a catalytic mechanism, in which MSA must enter the active site first, before acetyl-CoA can bind. After the reaction has taken place, formyl-CoA is released before acetoacetate can leave the active site. This hypothesis is further corroborated by the observation that acetoacetate is an inhibitor for the reaction with MSA and acetyl-CoA.

Earlier studies had proposed R254 as a catalytic base involved in the reaction (Supplementary Fig. 16)^28,29. However, our structures with bound CoA-ester prompted us to revisit the nature of the catalytic base and suggest a different mechanism for the reaction with MSA, in which E171 acts as the catalytic base (Fig. 2G, discussed in more detail in Supplementary Note 2, Supplementary Fig. 17). Replacing E171 by leucine abolished BKACE activity, while substitution by a chemically similar but shorter aspartate still showed basal activity (5% of the wild type; Supplementary Table 4), supporting the role of E171 in catalysis.

Modular implementation of the CORE cycle in vivo: Modules 2 + 4

We next aimed to demonstrate the CORE cycle in vivo. To implement the CORE cycle in E. coli, we used a modular strategy, in which parts of the pathway (i.e., “pathway modules”) are successively assembled and tested in growth-coupled selection strains^{38,39,40,41,42,43,44,45,46}. For the CORE cycle, we defined four modules (Fig. 3): Module 1 converts one molecule of acetyl-CoA and bicarbonate into MSA. Module 2 consists of the BKACE reaction and produces formyl-CoA and acetoacetate. Module 3 closes the CORE cycle by regenerating acetyl-CoA from acetoacetate. Lastly, module 4 enables growth-coupled selections by converting formate into the cellular building blocks formyl-tetrahydrofolate (THF), methylene-THF and serine (by condensation of methylene-THF with glycine).

Fig. 3: Subdivision of the CORE cycle into pathway modules facilitates stepwise implementation in *E. coli.*

We first constructed an E. coli strain that allows for growth-coupled detection of formate (i.e., the CORE cycle product). Guided by previous studies^{21,38,39,40,42,43,44,47,48}, we deleted serA and gcvTHP (encoding 3-phosphoglycerate dehydrogenase and the glycine cleavage system, respectively), creating a strain that is unable to synthesize serine as well as the intermediates of one-carbon metabolism (methyl-tetrahydrofolate (THF), methylene-THF and formyl-THF; jointly referred to as C1-THF in the following). In this deletion background, we chromosomally integrated three genes from Methylobacterium extorquens to enable formate-dependent biosynthesis of C1-THF and serine^39,42 (Fig. 4A): ftfL (formate-THF ligase), fchA (methenyl-THF cyclohydrolase), and mtdA (methylene-THF dehydrogenase). We additionally deleted the frmRAB operon (removing formaldehyde oxidation as a potential intracellular source of formate) as well as two endogenous alcohol dehydrogenases that are known to drain the CORE cycle intermediate MSA (∆ydfG ∆rutE^49,50,51,52). The resulting strain is referred to as “C1S-Aux” in the following.

**Fig. 4: Modules 2 and 4 are functional together in an *E. coli* selection strain.**

Notably, disruption of serine biosynthesis (∆serA) also abolished canonical glycine biosynthesis in the C1S-Aux strain. Although glycine could be generated from threonine cleavage (via ltaE or kbl-tdh), the corresponding genes are not constitutively expressed^42,53. Thus, we supplemented small amounts of glycine (2 mM) to minimal media in all experiments. In minimal medium containing a main carbon source (e.g. glycerol), growth of the C1S-Aux strain was strictly dependent on the supplementation of formate (and glycine), allowing robust detection of formate concentrations as low as 80 µM (Supplementary Fig. 18). Altogether, these experiments established the formate selection strain and verified CORE cycle module 4 in E. coli.

We next tested module 2, which comprised BKACE15 (Fig. 3). To provide MSA as the substrate for the BKACE reaction in vivo, we co-expressed a heterologous β-alanine:pyruvate transaminase (BPT; UniProt Q9I700⁵⁴), which allowed for the intracellular generation of MSA from β-alanine, when this compound was provided in the medium (Fig. 4A). Besides BKACE15, we also tested seven of the well-performing BKACE candidates identified in the in vitro screen, of which only BKACE137 achieved comparable growth rates (Fig. 4B). Growth was strictly β-alanine-dependent and correlated with the supplied β-alanine concentration, suggesting that module 2 of the CORE cycle was active (Fig. 4C). To independently confirm module 2 activity, we used stable isotope tracing with ¹³C₃-¹⁵N-β-alanine as substrate. These experiments showed almost exclusively single-labeled (M + 1) serine and methionine, confirming that the C1-derived atoms of these amino acids were indeed produced through modules 2 + 4 of the CORE cycle (Supplementary Fig. 19).

Modular implementation of the CORE cycle in vivo: Modules 1 + 2 + 4

Having demonstrated that modules 2 and 4 of the CORE cycle were functional in vivo, we sought to add more pathway modules towards assembling the entire cycle. We focused on reconstituting module 1 next. For the conversion of acetyl-CoA into malonyl-CoA, we relied on endogenous acetyl-CoA carboxylase activity (Fig. 5A), while for the subsequent reduction of malonyl-CoA to MSA, we expressed the C-terminal domain of the bi-functional MCR from C. aurantiacus³⁷ (see above).

Fig. 5: Modules 1, 2 and 4 of the CORE cycle operate robustly together in *E. coli.*

Simultaneous expression of MCR and BKACE15 (from the pCORE plasmid) rescued growth of the C1S-Aux selection strain. Notably, growth was observed on various tested carbon sources including glycerol, d-glucose and l-Lactate (Fig. 5B), indicating that the CORE cycle modules 1, 2 and 4 are robust across different steady-state concentrations of acetyl-CoA¹⁷. Control strains expressing only BKACE15 or MCR (i.e., lacking either module 1 or 2, respectively) failed to grow under the same conditions (Fig. 5B).

Notably, even though E. coli is known to grow on acetoacetate utilizing the native ato operon^52,55,56,57 (AtoDA and AtoB comprising the CORE cycle module 3), we could not observe growth of the C1S-Aux + pCORE strain with acetoacetate as the main carbon source. We observed growth only with externally provided formate, which rescues the selection (Fig. 5B). We thus sought to employ adaptive laboratory evolution (ALE) to establish the complete CORE cycle including module 3 in vivo.

Adaptive laboratory evolution enables operation of the complete CORE cycle

Operating the CORE cycle in vivo requires an intricate balancing of enzyme activities within the cycle and between overlapping pathways. This is particularly apparent for acetyl-CoA, which is a central metabolite in energy conservation (TCA cycle), fatty acid and amino acid metabolism, and also represents a branching point within the CORE cycle itself, because it is a substrate for the carboxylation reaction and BKACE. We hypothesized that a simple overexpression of enzymes was insufficient to balance fluxes, and thus had prevented growth of the C1S-Aux + pCORE strain on acetoacetate, so far.

Noting that the C1S-Aux + pCORE strain grew when provided with external formate (Fig. 5B), we designed a strategy to gradually evolve growth on acetoacetate with ALE (Fig. 6A, B). We first weaned the strain off its dependence on formate by passaging it via serial dilutions, while gradually decreasing the supplemented formate concentration from 200 µM to 0 µM (see “Methods”, Fig. 6C). After four passages (i.e., ~26 generations), one population slowly grew on acetoacetate in the absence of added formate. We isolated three random clones and confirmed their growth on acetoacetate. The most robustly growing clone, termed “Evo1”, reached a maximal growth rate of 0.063 ± 0.001 h⁻¹ (corresponding to a doubling time (T_d) of 11 h, Supplementary Fig. 20), but still showed much faster growth when formate was added to the medium.

**Fig. 6: ALE enables operation of the complete CORE cycle.**

To improve growth of this strain, we leveraged automated ALE. We continuously cultivated Evo1 in a GM3 automated evolution device^48,58,59 for 190 days (6.3 months; i.e., ~750 generations) and repeatedly diluted the culture with fresh acetoacetate medium when it reached an OD₆₀₀ of 0.4, thereby selecting for mutants with an increased growth rate (turbidostat regime). Over the course of this long-term evolution, the apparent growth rate of the population in the bioreactor improved ~4-fold from 0.046 h⁻¹ (doubling time (T_d) ~ 15.1 h) to 0.182 h⁻¹ (T_d ~ 3.8 h), indicating improved operation of the CORE cycle (Fig. 6D).

We isolated a random clone each from three different time-points of the ALE (Evo2, Evo3, Evo4) and compared their growth to the non-evolved “ancestor” (C1S-Aux + pCORE), as well as Evo1, from which ALE was started. As expected, clones Evo2-4 showed increased growth rates compared to Evo1 in acetoacetate medium, while the ancestor remained unable to grow in this condition (Fig. 6E). The fastest clone, Evo4, isolated from the final population, achieved a maximal growth rate of 0.18 h⁻¹ on acetoacetate (T_d = 3.9 h; Fig. 6E). Growth on acetoacetate was abolished upon curing the pCORE plasmid from clones Evo1-4, demonstrating that the heterologous CORE cycle enzymes were essential (Supplementary Fig. 21, further discussed below). To confirm that Evo3 and Evo4 were indeed using the CORE cycle, we used 1,3-¹³C₂-labeled acetoacetate as carbon source, which resulted in a serine and methionine labeling pattern that is consistent with a fully labeled C1-THF pool (Supplementary Fig. 22). Altogether, this data validated operation of the complete CORE cycle in the evolved E. coli strains.

Genetic basis of evolved growth via the CORE cycle

To understand the genetic adaptations along the ALE trajectory, we re-sequenced the genome and the plasmids of the non-evolved ancestor C1S-Aux and clones Evo1-4. Evo1, 2 and 3 contained nine distinct mutations in total, while Evo4 had accumulated 81 mutations (Supplementary Table 6), likely due to a transposon-mediated disruption of the mutT gene, which is known to increase the global mutation rate by more than thousandfold^52,60,61,62. Emergence of such hypermutator genotypes has been observed previously in ALE experiments, where it often serves as an “enabler” mutation^63,64,65,66.

To trace the appearance of this hypermutation genotype in our ALE, we isolated and sequenced additional random clones from the same Evo1-4 populations. This showed that the mutT mutation must have appeared between Evo2 and Evo3. Almost all clones of the Evo3 population already carried this mutation, which had completely taken over the Evo4 population (for all sequencing results, see Supplementary Data 1). Overall, we discovered 252 distinct mutations among the various clones, but decided to focus on two that were conserved in all clones and had affected two enzymes of the CORE cycle, MCR and AtoA.

Despite sequencing single clones, we found mixed sequencing reads for the MCR gene (encoded on the pCORE plasmid), indicating presence of both a wild-type and a mutated (i.e., truncated) MCR Y331* variant (Supplementary Table 6, Supplementary Data 1). We hypothesized that the mixed sequencing reads were caused by multimerized plasmids (concatemers)^67,68,69. We isolated the pCORE plasmids from Evo1-4 and confirmed the presence of concatemers in all evolved clones (Supplementary Fig. 23). Cell-extract assays showed that MCR activity was reduced between ~60–90% in clones carrying the concatemers (Supplementary Fig. 24). However, we found that concatemer formation was helpful, but not necessary for growth on acetoacetate, as follows: When transforming the non-evolved C1S-Aux with the isolated concatemer, this did not restore growth on acetoacetate. In contrast, when (re-)transforming a cured Evo3 strain with either the original pCORE plasmid, or a plasmid in which MCR expression was driven from a 10-fold weaker RBS^41,70 (“pCORE_rbsC”), or the concatemer (“pCORE_Evo3”), we observed growth in all cases. Yet, Evo3 strains transformed with pCORE_rbsC grew slightly better, indicating that a reduced MCR activity was helpful to operate the CORE cycle in vivo (Supplementary Fig. 21A).

In case of atoA, the gene had independently mutated (at least twice) during ALE, resulting in an AtoA I25L mutation in Evo1 and all derived clones (i.e., Evo2-4), while a parallel lineage (“Evo1a”) had acquired an AtoA A65S mutation (Supplementary Data 1). Homology modelling located I25 and A65 in the active site facing the acyl-moiety, indicating that these mutations might have altered substrate specificity or activity (Supplementary Fig. 25A). We tested the effect of these two mutations and the double variant (AtoA I25L A65S) on acetoacetate activation with acetyl-CoA and formyl-CoA (Supplementary Figs. 25B–D, Supplementary Table 7). While all variants exhibited lower activity compared to the WT, AtoA I25L (and the double variant) showed a 3- to 5-fold increased relative activity with formyl-CoA compared to acetyl-CoA, indicating that these variants had acquired the capability to activate acetoacetate more preferentially with formyl-CoA, thus integrating better with the CORE cycle (Fig. 6A).

The CORE cycle could enable synthetic autotrophy and enhance photosynthesis

Having demonstrated a functional CORE cycle in vivo, we finally analyzed its potential in improving or replacing natural CO₂ fixation (Fig. 1C, D). Using flux balance analysis⁷¹, we first evaluated the CORE cycle in combination with the Calvin-Benson-Bassham (CBB) cycle of photosynthesis. The CORE cycle can replace (or supplement) natural photorespiration^{22,72,73,74,75} and circumvent the wasteful decarboxylation of glycine by providing an alternative route to methylene-THF (Fig. 1D, Supplementary Fig. 26). This converts photorespiration into a process that assimilates CO₂ rather than releasing it, thus doubling its carbon efficiency from 75% to 150%^22,33,75,76. In terms of energetic efficiency, the CORE cycle requires 27% less ATP and 16% fewer reducing equivalents compared to native photorespiration and outperforms it by 25% in terms of 3PG yield (Supplementary Fig. 27, Supplementary Table 8).

We also assessed the CORE cycle as a CO₂-fixation pathway for synthetic autotrophy^{22,77,78,79,80,81}, for which we combined the CORE cycle with different downstream formate-assimilation pathways⁸². We modelled CO₂-fixation efficiency of the CORE cycle in combination with the serine cycle^42,44,83, the reductive glycine pathway⁴³, the formolase pathway^84,85, and the ribulose monophosphate (RuMP) cycle^86,87. This showed that all CORE cycle-based pathways outperformed the CBB cycle in terms of biomass yield (when considering RuBisCO oxygenation, Supplementary Fig. 28). Overall, these results showcase the flexibility of an oxygen-tolerant reduction-first strategy to support or replace natural CO₂ fixation pathways.

Discussion

In this work, we developed and realized the CORE cycle – an artificial, modular metabolic pathway for the conversion of CO₂ to formate. The CORE cycle allows CO₂-reduction at fully aerobic conditions and atmospheric CO₂ levels with NADPH as a reductant. We show that this metabolic route provides sufficient flux to satisfy the cellular demand for C1-units and serine in engineered E. coli strains, enabling a doubling time of roughly 4 h, which we achieved by a combination of rational metabolic engineering and ALE.

When coupled to downstream formate-assimilation pathways, the CORE cycle opens up new possibilities for CO₂-based bioproduction (Supplementary Fig. 28). While we demonstrate the successful production of cellular C1-units from CO₂ in this study, future efforts could open the way to more complex compounds, especially, when the CORE cycle is used in combination with recently developed synthetic pathways, such as the ATP-driven reduction of formate to formaldehyde^84,88 together with the homoserine cycle⁸⁹, the EuMP cycle⁹⁰, the FORCE pathway^91,92,93, or formolase⁸⁵.

Used as a CO₂-assimilating photorespiratory bypass, the CORE cycle carries the potential to enhance natural carbon fixation by providing an intracellular formate pool from CO₂, which could increase photosynthetic yield by 25% (Supplementary Fig. 27). In fact, it is tempting to speculate that such a “formate-assisted photorespiration” may occur in natural autotrophic, aerobic organisms which operate the CBB cycle in presence of reduced C1-compounds (e.g. methane, methanol, formate). In such organisms, these C1-compounds could serve not only as an energy source^94,95,96, but also as a source of C1-THF to support phosphoglycolate salvage⁹⁷, although this remains to be demonstrated.

We note that recently, the synthetic 4-hydroxy-2-oxobutanoic acid (HOB) cycle was presented, which in principle could also be considered as a CO₂-reduction pathway⁹⁸. The HOB cycle converts CO₂ into a THF-bound formaldehyde-moiety (methylene-THF) but has yet to be fully demonstrated in vivo (i.e., as a complete cycle). More importantly, this pathway does not produce a free C1 molecule, such as formate in case of the CORE cycle, making the HOB cycle less flexible in terms of downstream assimilation strategies (see above).

Overall, our study expands the natural solution space of “reduction-first”-based CO₂-fixation by an oxygen-tolerant alternative, opening up this highly efficient strategy for aerobic organisms, which provides multiple opportunities to enhance or replace natural and synthetic CO₂ fixation in the future.

Methods

Synthesis of chemical compounds

Malonate semialdehyde (MSA) was synthesized chemically⁹⁹, as follows: Alkaline hydrolysis of ethyl 3,3-diethoxypropionate (EDEOP, Sigma Aldrich) was followed by acid hydrolysis to yield MSA. 0.5 mL EDEOP were added to 7.5 mL of 0.5 M NaOH and incubated for 1 h at RT. Afterwards 5 mL 10% HCl were added and incubated for 1 h at RT. The solution was cooled on ice, 13 mL 1 M K_xH_yPO₄ buffer pH 6 were added and the pH was adjusted to pH 6. The solution was stored at −80 °C. The yield of MSA was determined enzymatically with full-length MCR from C. aurantiacus or 3-hydroxybutyrate dehydrogenase. The synthesis yield of MSA was estimated to be over 98%. Due to its instability, MSA was prepared freshly for each quantitative measurement.

Formyl-CoA was synthesized as follows, using slight modifications of established methods^100,101,102. Initially, formyl thiophenol was prepared by adding 5.8 mL (150 mmol, 3 eq.) of formic acid dropwise to 7.1 mL (75 mmol, 1.5 eq.) of acetic anhydride, followed by stirring at 25 °C for 2.5 h. Next, 61 µL (0.75 mmol, 0.015 eq.) of pyridine and 5.1 mL (50 mmol, 1 eq.) of thiophenol were added, and the mixture was stirred overnight. Impurities were removed using a rotary evaporator at 50 °C and 25 mbar. The remaining mixture was washed with cold brine, dried over MgSO₄, and distilled at 131 °C and 50–60 mbar to yield a clear oil, which was stored under nitrogen at −20 °C. Next, to synthesize formyl-CoA, 200 mg of CoA were dissolved in 2 mL of ice-cold 1 M KHCO₃ at pH 8.0 and degassed by shaking. Then, 0.4 mL of formyl thiophenol was added, and the mixture was shaken vigorously for 10 min. Cold diethyl ether was used to wash the product, removing phenol and formic acid. The pH was then lowered to below 4 by adding HCl, followed by two additional ether washes. The reaction product was purified by HPLC-MS¹⁰¹.

Acetyl-CoA was synthesized by symmetric anhydride synthesis¹⁰³. 200 mg CoA (0.25 mmol, 1 eq.) was mixed with 1.6 eq. of acetic anhydride (45 μL, 0.41 mmol) in 5 mL 0.5 M NaHCO₃ and stirred on ice for 45 min. Next, the pH of the solution was lowered to below 4 by adding formic acid and the reaction product was purified by HPLC-MS¹⁰³.

1,3-¹³C₂-acetoacetate was synthesized by alkaline hydrolysis of commercially available 1,3-¹³C₂-ethyl acetoacetate (Scientific Laboratory Supplies, Cat. No. 485640). A 1.3:1 ratio of NaOH and 1,3-¹³C₂-ethyl acetoacetate was mixed by dissolving 208 mg NaOH in 2 mL H₂O and adding the labeled compound. After 48 h incubation at room temperature and stirring conditions, the sample was frozen in liquid nitrogen, lyophilized and its concentration and purity validated by LC-MS analysis.

Protein production and purification

The collection of β-keto acid cleavage enzymes (BKACE) considered here was previously described²⁹. Briefly, the forward primers introduced a hexahistidine sequence in the proteins after the initial methionine for purification purposes¹⁰⁴. For each construction, the plasmid was transformed into BL21-CodonPlus (DE3)-RIPL cells.

For the production of cell lysates to screen the whole BKACE collection, corresponding plasmids were transformed into E. coli BL21-CodonPlus (DE3)-RIPL cells and grown in 96-well in 1.6 mL of Terrific Broth medium containing 0.5 M Sorbitol, 5 mM betaine and 100 µg/mL carbenicillin for 4.5 h at 37 °C. Isopropyl β-D-thiogalactopyranoside (IPTG) was added at a concentration of 500 μM to induce protein production, and the cells were further grown at 20 °C overnight. The cells were harvested by centrifugation and suspended in 300 µL of 50 mM Tris-HCl, pH 7.5 containing 10% glycerol, 1 mM Pefabloc SC and 0.2 µL of Lysonase™ Bioprocessing reagent (Merck NovagenR), and sonicated using a Branson 2510 sonication water bath. After centrifugation, the clarified lysate was analyzed by SDS-PAGE to check for recombinant protein production and stored at −80 °C. Protein concentration was determined by the Bradford assay with Bovine serum albumin as standard (Bio-Rad).

For the purification of BKACE homologs for in-depth characterization, the corresponding plasmid was transformed into chemically competent E. coli Arctic Express DE3. Cells were grown on lysogeny broth (Miller recipe) agar plates containing 34 µg/mL chloramphenicol an incubated over night at 37 °C. 2 L Terrific Broth medium (TB; 24 g/L yeast extract, 12 g/L tryptone, 4 mL/L glycerol, 17 mM KH₂PO₄, 72 mM K₂HPO₄) was inoculated from the agar plate and incubated at 37 °C and 140 rpm. At OD₆₀₀ = 0.4–0.6 protein expression was induced with 500 µM IPTG and cells were incubated over night at 16 °C. Cell harvesting at 8000 g and 4 °C for 12 min, lysis by French pressing at 18,000 psi and ultracentrifugation at 100,000 g was followed by His-Trap purification using an Äkta Start (GE Healthcare) linked to a HisTrap FF column (GE Healthcare). The purification buffer contained 50 mM HEPES pH 7.8, 500 mM KCl, 1 M proline and 1 mM ZnCl₂. The elution was done with 500 mM imidazole. Protein desalting occurred via gel-filtration chromatography using a HiLoad 16/600 Superdex 200 pg column (GE Healthcare) and a buffer containing 50 mM HEPES pH 7.8., 150 mM KCl, 1 M proline and 1 mM ZnCl₂. Desalting of protein samples for ICP-OES measurements was done by gel-filtration using the same buffer but without ZnCl₂. Protein quantification occurred by absorbance measurement at 280 nm. Protein purity was validated by SDS-PAGE using 15 µg of purified protein on a 4–20% Mini-Protean TGX Precast Protein Gel (Biorad).

Protein production and purification of Pcc_Me_D407I was done using an E. coli BL21 DE3 strain that contains a biotin ligase (BirA) from Methylobacterium extorquens and 6 L of TB containing additionally 2 µg/mL biotin. The purification buffer contained 50 mM HEPES pH 7.8 and 500 mM KCl and the desalting buffer contained 50 mM HEPES pH 7.8 and 150 mM KCl. All other steps remained the same as above.

For the production of all other proteins, the corresponding plasmid was transformed into chemically competent E. coli BL21 DE3. Cells were grown on lysogeny broth (Miller recipe) agar plates with appropriate antibiotics an incubated over night at 37 °C. 1 L TB medium was inoculated from the agar plate and incubated at 37 °C and 140 rpm. At OD₆₀₀ = 0.4–0.6 protein expression was induced with 500 µM IPTG and cells were incubated over night at 25 °C. Cell harvesting at 8000 g and 4 °C for 12 min, cell lysis using a Sonopuls GM200 sonicator (BANDELIN Electronic) at an amplitude of 50% with five consecutive cycles of 30 pulses over 60 s and ultracentrifugation at 100,000 g was followed by His-Trap purification using an Äkta Start (GE Healthcare) linked to a HisTrap FF column (GE Healthcare). The purification buffer contained 50 mM HEPES pH 7.8 and 500 mM KCl. The elution was done with 500 mM imidazole. Protein desalting occurred using a HiTrap Desalting column (GE Healthcare) and a buffer containing 50 mM HEPES pH 7.8. and 150 mM KCl. Protein quantification and purity validation occurred as described above.

Spectrophotometric assays

To measure NAD(P)H consumption or production over time, spectrophotometric assays with the Agilent Technologies Cary 60 UV-Vis device were done. All assays were carried out in a reaction volume of 200 µL using a quartz cuvette with 10 mm light path (Hellma Analytics) and the Kinetics software from Agilent.

To measure acetyl-CoA carboxylase (Pcc_Me_D407I) activity, a coupled enzyme assay with the full-length malonyl-CoA reductase (Mcr_Ca) from Chloroflexus aurantiacus was performed¹⁰⁵, as follows: 100 mM MOPS/KOH pH 7.8, 50 mM KHCO₃, 2 mM ATP, 0.3 mM NADPH, 5 mM MgCl₂, 2.8 mg/mL Mcr_Ca and 0.01−1 mg/mL Pcc_Me_D407I were mixed in a cuvette and incubated for 2 min at 37 °C. The reaction was started with 1 mM acetyl-CoA and absorption was measured over time at 340 nm.

To measure acetoacetate-CoA ligase (Acl_Sz/Acl_Sl) activity, a coupled enzyme assay with 3-hydroxybutyryl-CoA dehydrogenase (PhaB) from Ralstonia eutropha was performed. 100 mM MOPS/KOH pH 7.8, 0.6 mM NADH, 2.5 mM ATP, 2 mM acetoacetate, 5 mM MgCl₂, 0.075 mg/mL PhaB and 0.001–0.05 mg/mL Acl_Sz/Acl_Sl were mixed in a cuvette and incubated for 2 min at 37 °C. The reaction was started with 1 mM CoA and absorption was measured over time at 340 nm.

To measure malonyl-CoA reductase activity of purified proteins, 100 mM MOPS/KOH pH 7.8, 0.3 mM NADPH, and 0.08–0.32 mg/mL MCR were mixed in a cuvette and incubated for 2 min at 20 - 37 °C. The reaction was started with 1 mM malonyl-CoA and absorbance was measured over time at 340 nm. MCR of Sulfolobus tokodaii (Mcr_St) and the dissected C-terminal domain of Chloroflexus aurantiacus (MCR) were used. To measure malonyl-CoA reductase activity of clarified E. coli cell lysate, 0.04–1.7 mg/mL total proteins were used.

To measure β-ketothiolase (AtoB) activity, a coupled enzyme assay with propionaldehyde dehydrogenase (PduP) from Rhodopseudomonas palustris was performed. Acetoacetyl-CoA was produced in situ with Acetoacetate:CoA ligase (Acl_Sz) from Shinella zoogloeoides. 100 mM MOPS/KOH pH 7.8, 0.3 mM NADH, 2.5 mM ATP, 5 mM MgCl₂, 5 mM acetoacetate, 0.1 mg/mL Acl_Sz, 0.05 mg/mL PduP and 1.5–15 µg/mL AtoB were mixed in a cuvette and incubated for 2 min at 37 °C. The reaction was started with 2 mM CoA and absorbance was measured over time at 340 nm.

To measure the metal dependence of β-keto acid cleavage enzyme (BKACE15), a coupled enzyme assay was performed with malic dehydrogenase from porcine heart (MDH, Merck, CAS number 9001-64-3) and citrate synthase from porcine heart (CS, Merck, CAS number 9027-96-7). Oxaloacetate production by MDH is thermodynamically hindered and only occurs when oxaloacetate is drained by CS, requiring the co-substrate acetyl-CoA that is produced by the (reverse) BKACE reaction. 100 mM TRIS/HCl pH 7.5, 1 mM NAD⁺, 10 mM malate, 20 mM acetoacetate, 27.5 units MDH, 13.5 units CS, 0.1–0.5 mg/mL BKACE15, were mixed in a cuvette with either 5 mM ZnCl₂, 5 mM MgCl₂, 5 mM MnCl₂ or no metals and incubated for 2 min at 37 °C. The reaction was started with 0.5 mM formyl-CoA and absorbance was measured over time at 340 nm. BKACE15 activity was observed only upon (re-)addition of ZnCl₂.

LC-MS-based assays for BKACE activities

The β-keto acid cleavage enzyme (BKACE) collection screen was performed in recombinant cell-free extracts in 96 well-microplates at room temperature for 2 h. For each candidate enzyme roughly 12 µg of protein were incubated in 100 µL of 50 mM Tris-HCl (pH 7.5) containing 250 µM acetyl-CoA, 1.2 mM MSA. Due to the instability of acetoacetate in the ion source that prevents its reliable detection and quantitation, it was chemically derivatized with 2-Hydrazinoquinoline (2-HQ), to enhance chromatographic performance, stability and detectability by LC-MS¹⁰⁶. Specifically, product formation was tracked by detecting the loss of CO₂ from derivatized acetoacetate after CID fragmentation. 5 µL of the reaction were mixed with 100 µL of acetonitrile containing 50 µM 2-HQ and incubated for 1 h on ice. The derivatized sample was finally filtered on an AcroprepAdv Multi-well filter plate 0.2 µM WWPTFE (PALL). Experiments were conducted in duplicates.

For all BKACEs that reacted with MSA, proteins were purified and activities were quantified as follows: 3 µg of enzyme were incubated in 100 µL of 50 mM Tris-HCl (pH 7.5) containing 250 µM acetyl-CoA, 1.2 mM MSA. At different times (0, 30, 60 and 120 min) 5 µL of the reaction mixture were mixed with 100 µL of acetonitrile containing 50 µM 2-HQ and incubated for 1 h on ice. The derivatized sample was finally filtered on an AcroprepAdv Multi-well filter plate 0.2 µM WWPTFE (PALL). Experiments were conducted in duplicates. Accumulation of acetoacetate over time was linear during this period of time. Acetoacetate was quantified using a derivatized standard acetoacetate calibration range.

LC-MS analyses of the samples were conducted using a Dionex UltiMate TCC-3000RS chromatographic system (Thermo Fisher Scientific) coupled to a Velos Pro Orbitrap Elite mass spectrometer (Thermo Fisher Scientific) fitted with a heated electrospray ionization source (HESI) operating in the positive ionization mode. The ionization spray (IS) was set to +3.5 kV and the capillary temperature at 275 °C. Sheath gas and auxiliary gas flow rates were set at 60% arbitrary units (a.u.) and 44 a.u., respectively. The S-lens RF level was set to 55%. Data were acquired in raw files and processed with the Qualbrowser module of Xcalibur 2.2 (Thermo Fisher Scientific) to access to elemental compositions. Fragmentation experiments were performed under collision induced dissociation (CID) with a normalized collision energy (NCE) of 22%. Mass spectra were acquired over an m/z range from m/z 50 up to m/z 1000 with the mass resolution set to 60,000 FWHM at m/z 400.

Chromatographic separation was achieved on an XBridge C18 column (150 × 4.6 mm; 5 µm; Waters) thermostated at 40 °C. The mobile phase flow rate was set at 0.3 mL/min, and 10 μL were injected. Mobile phase A consisted of 0.05% aqueous acetic acid containing 2 mM ammonium acetate, and mobile phase B consisted of 95% aqueous acetonitrile containing 0.05% acetic acid and 2 mM ammonium acetate. The gradient started at 100% A for 2 min, followed by a linear gradient to 100% B for 10 min, and remained at 100% B for 10 min. The system returned to the initial solvent composition in 2 min and re-equilibrated under these conditions for 3 min.

In-depth kinetic characterization of BKACE candidates

Enzyme activity assays for in-depth characterization of selected BKACE candidates from the collection screen were conducted in an optimized assay with purified proteins. A reaction mix containing 100 mM MOPS/KOH pH 7.8, 4 mM MSA, 1 mM ZnCl₂ and 0.1 mg/mL BKACE was prepared in a volume of 160 µL and incubated at 37 °C for 1 min. The reaction was started by adding 1 mM acetyl-CoA and samples were quenched after 30, 60, 90 and 120 s in 10% formic acid. Quenched samples were analyzed by quantifying acetoacetate, formyl-CoA and acetyl-CoA via LC-MS as described below. Michaelis-Menten kinetics for MSA were conducted using the same assay and 0.5–50 mM MSA. Michaelis-Menten kinetics for acetyl-CoA were conducted using the same assay and 0.025–2 mM acetyl-CoA. Product inhibition by acetoacetate to determine IC₅₀ was measured using the same assay and 5 mM MSA, 1 mM acetyl-CoA and 0–20 mM acetoacetate. Relative activities were calculated by normalizing the data to the activity with 0 mM acetoacetate. Product inhibition by acetoacetate to investigate the type of inhibition was measured using the same LC-MS assay and formyl-CoA for detection. 2 mM acetyl-CoA, 0.5–4 mM MSA and 0–8 mM acetoacetate were used.

Quantitative determination of the acetyl-CoA, formyl-CoA and CoA was performed using a LC-MS/MS. The chromatographic separation was performed on an Agilent Infinity II 1290 HPLC system using a Kinetex EVO C18 column (150 × 2.1 mm, 3 μm particle size, 100 Å pore size, Phenomenex) connected to a guard column of similar specificity (20 × 2.1 mm, 3 μm particle size, Phenomoenex) at a constant flow rate of 0.25 mL/min with mobile phase A being 50 mM Ammonium Acetate in water at a pH of 8.1 and phase B being 100% methanol (Honeywell, Morristown, New Jersey, USA) at 40 °C. The injection volume was 2 µL. The mobile phase profile consisted of the following steps and linear gradients: 0–1 min constant at 2.5% B; 1–6 min from 2.5 to 95% B; 6–8 min constant at 95% B; 8–8.1 min from 95 to 2.5% B; 8.1 to 10 min constant at 2.5% B. An Agilent 6470 mass spectrometer was used in positive mode with an electrospray ionization source and the following conditions: ESI spray voltage 4500 V, nozzle voltage 1500 V, sheath gas 400 °C at 11 L/min, nebulizer pressure 30 psi and drying gas 250 °C at 11 L/min. Compounds were identified based on their mass transition and retention time compared to standards. Chromatograms were integrated using MassHunter software (Agilent, Santa Clara, CA, USA). Relative abundance was determined based on the peak area. Mass transitions, collision energies, Cell accelerator voltages and Dwell times have been optimized using chemically pure standards. Parameter settings of all targets are given in Supplementary Table 9.

Quantitative determination of Acetoacetate was performed using LC-MS/MS. The chromatographic separation was performed on an Agilent Infinity II 1290 HPLC system using a Kinetex EVO C18 column (150 × 2.1 mm, 3 μm particle size, 100 Å pore size, Phenomenex) connected to a guard column of similar specificity (20 × 2.1 mm, 3 μm particle size, Phenomoenex) at a constant flow rate of 0.2 mL/min with mobile phase A being 0.1% formic acid in water and phase B being 0.1% formic acid in methanol (Honeywell, Morristown, New Jersey, USA) at 25 °C. The injection volume was 0.5 µL. The profile of the mobile phase consisted of the following steps and linear gradients: 0–5 min constant at 0% B; 5–6 min from 0% to 100% B; 6–8 min constant at 100% B; 8–8.1 min from 100% to 0% B; 8.1–12 min constant at 0% B. An Agilent 6470 mass spectrometer was used in negative mode with an electrospray ionization source and the following conditions: ESI spray voltage 2000 V, nozzle voltage 500 V, sheath gas 300 °C at 11 L/min, nebulizer pressure 45 psi and drying gas 170 °C at 5 L/min. Compounds were identified based on their mass transition and retention time compared to standards. Chromatograms were integrated using MassHunter software (Agilent, Santa Clara, CA, USA). Absolute concentrations were determined based on an external Standard curve. Mass transitions, collision energies, Cell accelerator voltages and Dwell times have been optimized using chemically pure standards. Parameter settings of all targets are given in Supplementary Table 10.

Crystallization and structure determination of BKACE15

The sitting-drop vapor-diffusion method was used for crystallization at 16 °C. For the crystallization of the apo-enzyme purified BKACE15 (8.6 mg/mL) in 50 mM HEPES pH 7.8, 150 mM KCl, 1 M L-proline, and 1 mM ZnCl₂ was mixed in a 1:1 ratio with 25% (w/v) pentaerythritol propoxylate (17/8 PO/OH), 100 mM HEPES pH 7.5. The final size of the drops was 1 microliter. Prior to flash freezing the crystals in liquid nitrogen, the mother liquor was supplemented with 37% (w/v) pentaerythrol propoxylate (17/8 PO/OH).

Co-crystallization in the presence of malonate was performed by mixing BKACE15 (8.6 mg/mL) in 50 mM HEPES pH 7.8, 150 mM KCl, 1 M proline, and 1 mM ZnCl₂ in a 1:1 ratio with 20% w/v PEG3350, 200 mM di-sodium malonate pH 7.0. The final size of the drops was 1 µL. Prior to flash freezing the crystal in liquid nitrogen, the mother liquor was supplemented with 10 mM CoA and 40% (w/v) PEG200.

To acquire a crystal structure with bound acetyl-CoA and acetoacetate BKACE15 (8.6 mg/mL) in 50 mM HEPES pH 7.8, 150 mM KCl, 1 M L-proline, and 1 mM ZnCl₂ was mixed in a 1:1 ratio with 10% (v/v) pentaerythritol ethoxylate, 10% v/v 1-Butanol. The final size of the drops was 1 microliter. Prior to flash freezing the crystals in liquid nitrogen, the mother liquor was supplemented with 15 mM acetyl-CoA, 20 mM acetoacetate, and 20% (w/v) pentaerythrol propoxylate (17/8 PO/OH).

X-ray diffraction data (Supplementary Table 11) were collected at beamlines PETRA III P13 and P14 of DESY (Deutsches Elektronen-Synchrotron, Hamburg). Data were processed with the XDS¹⁰⁷ and CCP4 v.8.0 software suite¹⁰⁸. Structures were solved by molecular replacement using Phaser of the Phenix software package (v1.20.1)¹⁰⁹, built with Phenix.Autobuild and refined with Phenix.Refine. Additional modelling, manual refinement, and ligand fitting was done in Coot (v.0.9.8.3)¹¹⁰. Final positional and B-factor refinements, as well as water picking, were performed using Phenix.Refine. The metal dependence of BKACE was tested spectrophotometrically as described above. Structural models for the apo-enzyme, enzyme with bound CoA and malonate, as well as enzyme with bound acetyl-CoA and acetoacetate were deposited to the Protein Data Bank in Europe (PDBe) under PDB accession 8RIO, 8RIP, and 9HNF, respectively (Supplementary Table 11). Figures were made using PyMOL Molecular Graphics System (version 2.5.7; Schrödinger).

Metal analysis

Inductively coupled plasma optical emission spectroscopy (ICP-OES) was used to identify metals bound to BKACE15. For sample preparation, 236 µg of BKACE15 were dissolved in 0.5 mL of trace metal grade concentrated nitric acid and incubated for 12 h at 25 °C. Subsequently, the samples were boiled for 2 h at 70 °C before they were diluted 17-fold in distilled water. The metal content was analyzed with a 720/725 ICP-OES device (Agilent Technologies) and a λ = 213.857 nm for zinc. The device was operated with ICP Expert v4.1.0 software (Agilent Technologies). Zinc was quantified using ICP multi-element standard solution IV (Merck) as a standard. The results were plotted using GraphPad Prism v10.

Structural modelling and analysis

Structural investigation of atoA mutations was performed via homology modelling using SWISS-MODEL¹¹¹. As a template, the structure of succinyl-CoA:3-ketoacid CoA transferase from pig heart (PDB 3OXO, 45% amino acid identity) was used, which had a CoA bound that was covalently linked to an active site glutamate.

Strains and genomic modifications

All strains used in this study are listed in Supplementary Table 12. All selection strains are based on E. coli SIJ488¹¹², a derivative of E. coli MG1655. The SIJ488 strain is equipped with inducible genes for λ-Red recombineering (Red recombinase system; flippase) integrated into its chromosome to increase ease-of-use for multiple genomic modifications. Using a published strain as a starting point (Supplementary Table 12), the additional gene deletions were performed by λ-Red recombineering¹¹³, using linear dsDNA fragments produced by PCR of an antibiotic resistance cassette (Cm^R or Km^R) with primers carrying 50 bp overhangs serving as homologous regions flanking the target gene (for oligonucleotides, see Supplementary Table 13). For deletion of rutE, the Cm^R resistance cassette from pKD3¹¹³ was amplified. For deletion of ydfG, the Km^R resistance was amplified from the template “FRT-PGK-gb2-neo-FRT” (Cat. No. A002, Gene Bridges GmbH, Germany). The PCR product (400 ng) was electroporated into the desired strain, using cells harvested from a mid-log phase culture in which recombination proteins had been induced with 15 mM L-Arabinose for 45 min at OD₆₀₀ ~ 0.3. Strains were verified as described below.

For chromosomal integration of the heterologous genes (required for CORE cycle module 4, i.e. formate assimilation/C1-THF metabolism), we used a previously described genome insertion protocol⁴¹ and a published construct containing the desired genes (“pDM4:SS9-C₁M”)³⁹. In brief, a non-replicative plasmid (pDM4, R6K ori) was introduced into the recipient strain via conjugation from an E. coli ST18 donor strain. Antibiotic resistance was used to select for chromosomal insertion via native homologous recombination based on 600 bp homology regions, with subsequent levansucrase (sacB) counter-selection. Serendipitously, a clone was identified in which the complete operon had inserted not in the previously used genomic locus (“SS9”^39,114), but instead, in a non-coding region in-between nhaR and rpsT, likely due to unintended homology of this region with the pDM4 backbone. The insertion site was confirmed by PCR, Sanger sequencing and short-read sequencing. This clone was used for all further steps.

Strains with the desired deletions or chromosomal insertion were selected by plating on appropriate antibiotics (kanamycin or chloramphenicol) and confirmed by determining the size of the respective genomic locus via PCR (oligonucleotide sequences are shown in Supplementary Table 13). For removal of resistance markers flanked by FRT sites, flippase was induced in growing cultures at OD₆₀₀ ~ 0.2 by adding 50 mM L-rhamnose, followed by cultivating at 30 °C for ~4 h. Loss of the antibiotic resistance was confirmed by identifying individual colonies that required absence of the respective antibiotic, and by PCR of the genomic locus (oligonucleotide sequences are shown in Supplementary Table 13). Plasmids were transformed into selection strains by electroporation. Presence of the plasmid was verified by antibiotic resistance and PCR.

Cloning of expression constructs for protein purification

All plasmids used in this study are listed in Supplementary Data 2. Cloning of expression constructs for Acl_Sl, Acl_Sz, MCR and all BKACE-variants (for protein purification) was done using modular cloning and the “Marburg Collection”¹¹⁵. Target genes were either synthesized or PCR amplified from the host’s genome using Q5 polymerase from NEB and the dedicated protocol. Golden gate cloning was followed by DpnI digestion, transformation into E. coli NEB Turbo and colony picking to yield the correct construct that was validated by sequencing.

Cloning of the expression construct for AtoDA was done by Golden gate cloning of synthesized genes into pET28a, followed by DpnI digestion, transformation into E. coli NEB Turbo and colony picking to yield the correct construct that was validated by sequencing.

Introduction of point mutations to construct mutant variants of BKACE15 and AtoA was done by oligonucleotide-based site-directed mutagenesis¹¹⁶. pME_B_0_04_0055_BKACE15 and pTE3156 were used as templates for mutagenesis. The PCR was followed by DpnI digest and a PCR clean-up. The cloning mixture was transformed into E. coli NEB Turbo and colonies picked for sequence validation.

For transformation of cloned constructs, various amounts of DNA were added to 50 µL of chemically competent E. coli DH5α or E. coli NEB Turbo and incubated on ice for 30 min. A heat shock at 42 °C for 45 s was followed by 10 min of incubation on ice. 600 µL S.O.C. medium (Invitrogen) was added, and cells were incubated for 1 h at 37 °C while shaking at 200 rpm. Finally, cells were transformed on LB plates with selection marker.

Plasmid construction for in vivo testing of CORE cycle modules

Cloning was performed in E. coli DH5α. All plasmids used in this study are listed in Supplementary Data 2. Heterologous genes were codon-optimized for E. coli K−12 and synthesized by TWIST Biosciences. All heterologous genes for in vivo testing were expressed with an N-terminal 6xHisTag without a linker, to provide a consistent sequence immediately downstream of the ribosomal binding site (RBS), aiming to reduce sequence-dependent differences in translation initiation rate⁷⁰.

Genes were cloned into a pZASS backbone (p15A ori, Strep^R)⁴¹ containing a constitutive promoter (“pgi#20”)¹¹⁷. A medium-strength ribosome binding site (RBS “C”: aagttaagaggcaaga⁷⁰) was used for all BKACE candidates and BPT, while either RBS “C” or an RBS with a higher predicted translation initiation rate (RBS “B”: aacaaaatgaggaggtactgag⁷⁰) was used upstream of MCR (see Supplementary Data 2 for a list of plasmids used in this study, and Supplementary Data 3 for sequence maps).

Growth media and growth characterization

Lysogeny broth (LB) media was used for routine culturing (1% NaCl, 0.5% yeast extract and 1% tryptone), supplemented with appropriate antibiotics for maintenance of plasmids. M9 minimal medium (Sigma-Aldrich, Cat. No. M6030) was used for all growth experiments, supplemented with trace elements (134 μM EDTA, 13 μM FeCl₃, 6.2 μM ZnCl₂, 0.76 μM CuCl₂, 0.42 μM CoCl₂, 1.62 μM H₃BO₃, 0.081 μM MnCl₂). Carbon sources were added as indicated in the main text, e.g. glycerol (20 mM), glucose (10 mM/20 mM), sodium l-lactate (20 mM), sodium formate. Lithium acetoacetate (Sigma-Aldrich, Cat. No. A8509) was used as source of acetoacetate. Sodium 1,3-¹³C-acetoacetate was used for ¹³C-labeling (for synthesis, see above). Media containing acetoacetate were freshly prepared or stored at −20 °C to avoid variance in acetoacetate concentrations due to its long-term instability in aqueous solutions¹¹⁸.

Minimal media containing formate (5 mM) were used to preculture strains, enabling CORE cycle-independent growth. Precultures were grown in 3 mL medium in 10 mL glass tubes under vigorous shaking at 37 °C. Antibiotics were added to precultures, when appropriate, but omitted for growth experiments. When acetoacetate was used as main carbon source in the final growth experiment, the preculture was performed in M9 + lithium acetoacetate (20 mM) + glycine (2 mM) + formate (5 mM) and streptomycin (where appropriate for maintenance of a plasmid). For all other experiments, the preculture was grown with glycerol as default carbon source, i.e., M9 + glycerol (20 mM) + glycine (2 mM) + formate (5 mM) and streptomycin (where appropriate). Streptomycin was omitted for untransformed or cured strains lacking a plasmid.

Cells from the preculture were pelleted and washed thrice in M9 medium without carbon source, prior to inoculation in 96-well plates (Nunclon Delta Surface, Thermo Scientific, Dreieich, Germany) at a starting OD₆₀₀ of 0.02. Wells were filled with 150 µL culture and covered with 50 µL mineral oil (Merck, Darmstadt, Germany) to avoid evaporation while allowing gas exchange. Aerobic growth was monitored in technical replicates (as indicated) at 37 °C in a BioTek Epoch 2 Microplate Spectrophotometer (BioTek, Bad Friedrichshall, Germany) by absorbance measurements (600 nm) of each well every ~10 min with intermittent orbital and linear shaking. Blank measurements were subtracted and OD₆₀₀ measurements were normalized to OD₆₀₀ values corresponding to a path length of 1 cm (“cuvette OD”) by multiplying with a factor of 4.35, as established via a calibration curve for the instrument.

Maximal growth rates (and correspondingly, doubling times) were estimated for the mid-log phase using a custom MATLAB script.

¹³C isotopic labelling of proteinogenic amino acids

For isotope tracing, cells were cultured as triplicates (three separate cultures of the same clonal strain) in 14 mL glass culture tubes with 3 mL M9 medium containing the carbon sources. Cultures were inoculated to an OD₆₀₀ 0.01 and grown at 37 °C until an OD in late exponential phase, between 0.6–1.0. Then, a cell amount equivalent to 1 mL of culture with OD₆₀₀ 1 was pelleted (~10⁹ cells), washed once with ddH₂O and hydrolyzed in 1 mL hydrochloric acid (6 M) at 95 °C for 24 h. Subsequently, the acid was evaporated by heating at 95 °C under an air stream, and the hydrolyzed biomass was resuspended in 1 mL ddH₂O.

Relative quantification of isotopologues for amino acids of interest was performed using ultra-performance liquid chromatography high-resolution mass spectrometry (LC-MS). The chromatographic separation was performed on a Thermo Scientific Vanquish HPLC System using a ZicHILIC SeQuant column (150 × 2.1 mm, 3.5 μm particle size, 100 Å pore size) connected to a ZicHILIC guard column (20 × 2.1 mm, 5 μm particle size) (Merck KGaA, Darmstadt, Germany) at a constant flow rate of 0.3 mL/min with mobile phase A being 0.1% formic acid in 99:1 water:acetonitrile (Honeywell, Morristown, New Jersey, USA) and phase B being 0.1% formic acid 99:1 acetonitrile:water (Honeywell, Morristown, New Jersey, USA) at 25 °C. The injection volume was 1 µL. The mobile phase profile consisted of the following steps and linear gradients: 0–8 min from 80 to 60% B; 8–10 min from 60 to 10% B; 10–12 min constant at 10% B; 12–12.1 min from 10 to 80% B; 12.1 to 15 min constant at 80% B. A Thermo Scientific ID-X Orbitrap mass spectrometer was used in positive mode with a high-temperature electrospray ionization (H-ESI) source and the following conditions: H-ESI spray voltage at 3500 V, sheath gas at 50 arbitrary units, auxiliary gas at 10 arbitrary units, sweep gas at 1 arbitrary units, ion transfer tube temperature at 325 °C, vaporizer temperature at 350 °C. Detection was performed in full scan mode using the Orbitrap mass analyzer at a mass resolution of 60 000 in the mass range 50 - 250 (m/z). Extracted ion chromatograms of the [M + H]⁺ forms were integrated using the TraceFinder software (Thermo Scientific) applying a mass tolerance of 5 ppm. Amino acid standards (Merck KGaA, Darmstadt, Germany) were analyzed under the same conditions in order to determine expected retention times.

Isolation of mutants capable of growing via the complete CORE cycle

Initial adaptive evolution was performed by manual serial passaging in 4 mL cultures of “acetoacetate medium” (M9 + 20 mM lithium acetoacetate+2 mM glycine+varying sodium formate concentrations). Initially, acetoacetate medium containing 200 µM sodium formate was inoculated with C1S-Aux+pCORE to an OD₆₀₀ of 0.01. Once growth had reached a stationary plateau above OD 0.5, the population was passaged at a dilution of 1:100 into a new culture of 4 mL, with a quarter of the previous formate concentration. This process was repeated twice, at which point the formate concentration in the medium had been decreased to 12.5 µM. From this passage, cells were transferred to three cultures lacking formate entirely. One of these cultures showed visible growth within 14 days and individual colonies (denoted Evo1, Evo1a, Evo1b, Evo1c in Supplementary Data 1) were isolated on solid “acetoacetate medium” (see above, with added 1.2% ultra-pure agarose). Growth of these clones was assayed in liquid “acetoacetate medium” lacking formate, identifying Evo1 as the best-performing isolate showing reproducible growth, thus used for further ALE.

Long-term continuous culture evolution

Growth on acetoacetate as acetyl-CoA precursor of Evo1 bacteria (evolved from the C1S-Aux+pCORE strain by serial dilution) to close the CORE cycle was accelerated in a turbidostat conducted in a GM3 automated continuous culture device^48,58,59. An Evo1 preculture (16 mL acetoacetate medium: M9 + 20 mM lithium acetoacetate+2 mM glycine, pH = 7.2) was injected in a GM3 growth chamber and grown at 37 °C. Every 10 min, the optical density of the culture was automatically measured and compared to a fixed threshold (OD₆₀₀ value of 0.4). When the measured OD exceeded the threshold, a 10% dilution pulse (1.68 mL) of fresh acetoacetate medium was injected by the fluidic system into the culture and the same volume of used culture was discarded. The dilutions ensured that the biomass in the vessel remained constant and that the bacteria grew at their maximal growth rate. Biofilm formation at the inner side of the growth chamber was counteracted by regular cycles of washing with 5 N NaOH and subsequent rinsing of the chamber with H₂O, thus enabling long term culturing periods. Once a week, a 1 mL sample of the culture was withdrawn and stored upon addition of DMSO (10% final) at −80 °C to constitute a stock of culture intermediates, from which evolved clones were isolated (e.g. Evo2/3/4).

Whole genome re-sequencing

Genomic DNA was extracted using the NucleoSpin Microbial DNA purification kit (Macherey-Nagel, Germany), using cells harvested from an overnight culture in LB medium supplied with Streptomycin (to maintain the pCORE plasmid). Library preparation and short-read sequencing were performed by Novogene Inc. (Cambridge, UK). In brief, genomic DNA was randomly sheared into short fragments. The obtained fragments were end-repaired, A-tailed and further ligated with Illumina adapters. Fragments containing adapters were PCR-amplified, size-selected, and purified. Sequencing was performed on an Illumina NovaSeq X Plus platform to obtain 150 bp paired-end reads.

Mapping of reads and identification of mutations was performed using breseq v0.38.3¹¹⁹ with two reference sequences: (1) the pCORE plasmid, and (2) a manually adjusted reference sequence based on the genome of E.coli MG1655 (GenBank accession NC_000913)¹²⁰. The reference sequence used here contains several modifications compared to NC_000913, to reflect changes made during deletions and insertions in the genome (for a list of all sequence adjustments, see Supplementary Data 4).

Plasmid curing

Plasmids were cured from isolated, evolved clones by propagation in non-selective minimal medium lacking antibiotics and supplemented with formate to enable loss of pCORE (M9 + 20 mM lithium acetoacetate + 2 mM glycine + 5 mM sodium formate). Clones which had lost the plasmid were identified by individually streaking on LB with and without Streptomycin. Plasmid loss was further verified by absence of a PCR product when amplifying with primers P01 and P02 (Supplementary Table 13).

Lysate-based MCR activity measurements

Strains were grown to late exponential phase (OD 1.0) in 200 mL Terrific Broth (TB) medium with streptomycin, under vigorous shaking. Rich medium was chosen to enable comparison of the evolved strains with their non-evolved ancestor and plasmid-free C1S-Aux strain, both of which are incapable of fully CORE cycle-dependent growth. Cells from 200 mL culture were harvested by centrifugation, lysed in 50 mM HEPES pH 7.8, 150 mM KCl buffer via homogenization at 22,500 psi and cleared by ultracentrifugation at 100,000 g. MCR-activity was determined in a spectrophotometric assay as described above. Protein concentrations in clarified lysates were quantified using the Bradford assay (Quick Start Bradford 1x Dye reagent, BioRad). For higher sensitivity, absorbance was measured as 590/490 nm ratio¹²¹.

Flux balance analysis

To compare the CORE cycle variants to native plant photorespiration and previously proposed photorespiration bypasses, we performed stoichiometric modeling by applying flux balance analysis (FBA) with COBRApy (v0.20.0)¹²², following a framework we used previously³³. In brief, for each pathway, we calculated the consumption of ATP, NAD(P)H and reduced ferredoxins, as well as the required turns of the CBB cycle (including RuBisCO) to produce one unit of 3-phosphoglycerate. For this purpose, we constructed a simplified metabolic model consisting of the CBB cycle, specific reactions of each considered photorespiratory pathway, and generic/artificial cofactor regeneration and interconversion reactions (e.g., ADP + inorganic phosphate → ATP; NAD⁺ → NADH and NADH + NADP⁺ → NAD⁺ + NADPH; ATP + AMP → 2 ADP; oxidized ferredoxin → reduced ferredoxin). We note that transfer of electrons from NADH to NADPH, and vice-versa, is freely possible in this model without any ATP-investment via a generic transhydrogenase reaction (simulating photosynthetic production of NADPH). We assumed a RuBisCO carboxylation-to-oxygenation ratio of 3:1 (i.e. 25% of all RuBisCO reactions being oxygenations)^74,123,124. To compare the yield of all pathways, we calculated their total required ATP-equivalents to produce 3PG from CO₂, by using the conversions 1 NAD(P)H = 2.5 ATP^125,126 and 2 reduced ferredoxins = 1 NAD(P)H. The full code and the models can be found at the online repository Edmond (https://doi.org/10.17617/3.V8KJFV) within the “cost_comparison” directory. Model results are also provided in Supplementary Data 5.

FBA was also conducted to estimate the potential of the CORE cycle for synthetic autotrophy and CO₂-based bioproduction. We extended the CORE cycle with various formate assimilation pathways (see main text). We integrated the corresponding non-native reactions into the most updated E. coli genome-scale metabolic model (iML1515)¹²⁷, with several curations and changes (see full code linked below). We used H2 as a generic electron/energy source (via hydrogenase: H₂ + NAD⁺ → NADH + H⁺) and computed autotrophic yields of biomass and 12 bioproduction precursors¹²⁸ from CO₂. We used the CBB cycle (assuming RuBisCO carboxylation-to-oxygenation ratio of 3:1) as the benchmark to calculate the relative yields. The full script and models, including the individual changes, can be found at the online repository Edmond (https://doi.org/10.17617/3.V8KJFV) within the “bioproduction” directory.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

Structural models for the BKACE15 apo-enzyme, enzyme with bound CoA and malonate, as well as enzyme with bound acetyl-CoA and acetoacetate were deposited to the Protein Data Bank in Europe under PDB accession 8RIO, 8RIP, and 9HNF, respectively. Raw data from next-generation sequencing has been deposited in the Sequence Read Archive (SRA) under accession PRJNA1124540. Source data are provided with this paper.

Code availability

The code used for flux balance analysis has been deposited at Edmond [https://doi.org/10.17617/3.V8KJFV]. The code and reference sequences used to analyze sequencing data via breseq are also available from Edmond [https://doi.org/10.17617/3.NKWOPO].

References

Bar-Even, A., Noor, E. & Milo, R. A survey of carbon fixation pathways through a quantitative lens. J. Exp. Bot. 63, 2325–2342 (2012).
Article CAS PubMed MATH Google Scholar
Cotton, C. A., Edlich-Muth, C. & Bar-Even, A. Reinforcing carbon fixation: CO₂ reduction replacing and supporting carboxylation. Curr. Opin. Biotechnol. 49, 49–56 (2018).
Article CAS PubMed Google Scholar
Mall, A. et al. Reversibility of citrate synthase allows autotrophic growth of a thermophilic bacterium. Science 359, 563–567 (2018).
Article ADS CAS PubMed MATH Google Scholar
Steffens, L. et al. High CO₂ levels drive the TCA cycle backwards towards autotrophy. Nature 592, 784–788 (2021).
Article ADS CAS PubMed MATH Google Scholar
Sánchez-Andrea, I. et al. The reductive glycine pathway allows autotrophic growth of Desulfovibrio desulfuricans. Nat. Commun. 11, 5090 (2020).
Article ADS PubMed PubMed Central MATH Google Scholar
Berg, I. A. Ecological aspects of the distribution of different autotrophic CO2 fixation pathways. Appl Environ. Microbiol 77, 1925–1936 (2011).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Schuchmann, K. & Müller, V. Autotrophy at the thermodynamic limit of life: a model for energy conservation in acetogenic bacteria. Nat. Rev. Microbiol 12, 809–821 (2014).
Article CAS PubMed MATH Google Scholar
Fackler, N. et al. Stepping on the gas to a circular economy: accelerating development of carbon-negative chemical production from gas fermentation. Annu Rev. Chem. Biomol. Eng. 12, 439–470 (2021).
Article PubMed MATH Google Scholar
Flamholz, A., Noor, E., Bar-Even, A. & Milo, R. eQuilibrator-the biochemical thermodynamics calculator. Nucleic Acids Res. 40, D770–D775 (2012).
Article CAS PubMed Google Scholar
Thauer, R. K., Jungermann, K. & Decker, K. Energy conservation in chemotrophic anaerobic bacteria. Bacteriol. Rev. 41, 100–180 (1977).
Article CAS PubMed PubMed Central MATH Google Scholar
Andersen, K. B. & von Meyenburg, K. Charges of nicotinamide adenine nucleotides and adenylate energy charge as regulatory parameters of the metabolism in Escherichia coli. J. Biol. Chem. 252, 4151–4156 (1977).
Article CAS PubMed MATH Google Scholar
Lilius, E. M., Multanen, V. M. & Toivonen, V. Quantitative extraction and estimation of intracellular nicotinamide nucleotides of Escherichia coli. Anal. Biochem 99, 22–27 (1979).
Article CAS PubMed Google Scholar
Bautista, J., Satrústegui, J. & Machado, A. Evidence suggesting that the NADPH/NADP ratio modulates the splitting of the isocitrate flux between the glyoxylic and tricarboxylic acid cycles, in Escherichia coli. FEBS Lett. 105, 333–336 (1979).
Article CAS PubMed Google Scholar
Fuhrer, T. & Sauer, U. Different biochemical mechanisms ensure network-wide balancing of reducing equivalents in microbial metabolism. J. Bacteriol. 191, 2112–2121 (2009).
Article CAS PubMed PubMed Central MATH Google Scholar
Gerosa, L. et al. Pseudo-transition analysis identifies the key regulators of dynamic metabolic adaptations from steady-state data. Cell Syst. 1, 270–282 (2015).
Article CAS PubMed MATH Google Scholar
de Graef, M. R., Alexeeva, S., Snoep, J. L. & Teixeira de Mattos, M. J. The steady-state internal redox state (NADH/NAD) reflects the external redox state and is correlated with catabolic adaptation in Escherichia coli. J. Bacteriol. 181, 2351–2357 (1999).
Article PubMed PubMed Central Google Scholar
Bennett, B. D. et al. Absolute metabolite concentrations and implied enzyme active site occupancy in Escherichia coli. Nat. Chem. Biol. 5, 593–599 (2009).
Article CAS PubMed PubMed Central MATH Google Scholar
Noor, E. et al. Pathway thermodynamics highlights kinetic obstacles in central metabolism. PLoS Comput Biol. 10, e1003483 (2014).
Article PubMed PubMed Central Google Scholar
Noor, E., Flamholz, A., Liebermeister, W., Bar-Even, A. & Milo, R. A note on the kinetics of enzyme action: a decomposition that highlights thermodynamic effects. FEBS Lett. 587, 2772–2777 (2013).
Article CAS PubMed Google Scholar
Noor, E. et al. The protein cost of metabolic fluxes: prediction from enzymatic rate laws and cost minimization. PLoS Comput Biol. 12, e1005167 (2016).
Article PubMed PubMed Central Google Scholar
Bruinsma, L., Wenk, S., Claassens, N. J. & Martins Dos Santos, V. A. P. Paving the way for synthetic C1 - metabolism in pseudomonas putida through the reductive glycine pathway. Metab. Eng. 76, 215–224 (2023).
Article CAS PubMed Google Scholar
Bar-Even, A. Daring metabolic designs for enhanced plant carbon fixation. Plant Sci. 273, 71–83 (2018).
Article CAS PubMed MATH Google Scholar
Banerjee, R. & Ragsdale, S. W. The many faces of vitamin B12: catalysis by cobalamin-dependent enzymes. Annu Rev. Biochem 72, 209–247 (2003).
Article CAS PubMed MATH Google Scholar
Smith, A. G., Croft, M. T., Moulin, M. & Webb, M. E. Plants need their vitamins too. Curr. Opin. Plant Biol. 10, 266–275 (2007).
Article CAS PubMed Google Scholar
Erb, T. J., Jones, P. R. & Bar-Even, A. Synthetic metabolism: metabolic engineering meets enzyme design. Curr. Opin. Chem. Biol. 37, 56–62 (2017).
Article CAS PubMed PubMed Central Google Scholar
Khersonsky, O. & Tawfik, D. S. Enzyme promiscuity: a mechanistic and evolutionary perspective. Annu Rev. Biochem 79, 471–505 (2010).
Article CAS PubMed MATH Google Scholar
Chen, K. & Arnold, F. H. Engineering new catalytic activities in enzymes. Nat. Catal. 3, 203–213 (2020).
Article CAS MATH Google Scholar
Bellinzoni, M. et al. 3-Keto-5-aminohexanoate cleavage enzyme: a common fold for an uncommon Claisen-type condensation. J. Biol. Chem. 286, 27399–27405 (2011).
Article CAS PubMed PubMed Central Google Scholar
Bastard, K. et al. Revealing the hidden functional diversity of an enzyme family. Nat. Chem. Biol. 10, 42–49 (2014).
Article CAS PubMed MATH Google Scholar
Sly, W. S. & Stadtman, E. R. Formate metabolism. I. Formyl coenzyme a, an intermediate in the formate-dependent decomposition of acetyl phosphate in clostridium kluyveri. J. Biol. Chem. 238, 2632–2638 (1963).
Article CAS PubMed Google Scholar
Croes, K., Van Veldhoven, P. P., Mannaerts, G. P. & Casteels, M. Production of formyl-CoA during peroxisomal alpha-oxidation of 3-methyl-branched fatty acids. FEBS Lett. 407, 197–200 (1997).
Article CAS PubMed Google Scholar
Chou, A., Clomburg, J. M., Qian, S. & Gonzalez, R. 2-Hydroxyacyl-CoA lyase catalyzes acyloin condensation for one-carbon bioconversion. Nat. Chem. Biol. 15, 900–906 (2019).
Article CAS PubMed Google Scholar
Scheffen, M. et al. A new-to-nature carboxylation module to improve natural and synthetic CO₂ fixation. Nat. Catal. 4, 105–115 (2021).
Article CAS Google Scholar
Son, H. F. et al. Structural insight into bi-functional malonyl-CoA reductase. Environ. Microbiol 22, 752–765 (2020).
Article CAS PubMed Google Scholar
Hügler, M., Menendez, C., Schägger, H. & Fuchs, G. Malonyl-Coenzyme A reductase from Chloroflexus aurantiacus, a key enzyme of the 3-hydroxypropionate cycle for autotrophic CO₂ fixation. J. Bacteriol. 184, 2404–2410 (2002).
Article PubMed PubMed Central Google Scholar
Alber, B. et al. Malonyl-Coenzyme A reductase in the modified 3-hydroxypropionate cycle for autotrophic carbon fixation in archaeal Metallosphaera and Sulfolobus spp. J. Bacteriol. 188, 8551–8559 (2006).
Article CAS PubMed PubMed Central MATH Google Scholar
Liu, C., Wang, Q., Xian, M., Ding, Y. & Zhao, G. Dissection of Malonyl-Coenzyme A reductase of Chloroflexus aurantiacus results in enzyme activity improvement. PLoS One 8, e75554 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Yishai, O., Bouzon, M., Döring, V. & Bar-Even, A. In vivo assimilation of one-carbon via a synthetic reductive glycine pathway in Escherichia coli. ACS Synth. Biol. 7, 2023–2028 (2018).
Article CAS PubMed Google Scholar
Kim, S. et al. Growth of E. coli on formate and methanol via the reductive glycine pathway. Nat. Chem. Biol. 16, 538–545 (2020).
Article CAS PubMed MATH Google Scholar
Claassens, N. J., He, H. & Bar-Even, A. Synthetic methanol and formate assimilation via modular engineering and selection strategies. Curr. Issues Mol. Biol. 33, 237–248 (2019).
Article PubMed Google Scholar
Wenk, S., Yishai, O., Lindner, S. N. & Bar-Even, A. An engineering approach for rewiring microbial metabolism. Methods Enzymol. 608, 329–367 (2018).
Article CAS PubMed Google Scholar
Wenk, S. et al. Evolution-assisted engineering of E. coli enables growth on formic acid at ambient CO₂ via the Serine Threonine Cycle. Metab. Eng. 88, 14–24 (2025).
Article CAS PubMed MATH Google Scholar
Claassens, N. J. et al. Engineering the reductive glycine pathway: a promising synthetic metabolism approach for C1-assimilation. Adv. Biochem Eng. Biotechnol. 180, 299–350 (2022).
CAS PubMed MATH Google Scholar
Yu, H. & Liao, J. C. A modified serine cycle in Escherichia coli coverts methanol and CO₂ to two-carbon compounds. Nat. Commun. 9, 3992 (2018).
Article ADS PubMed PubMed Central MATH Google Scholar
Luo, S. et al. Construction and modular implementation of the THETA cycle for synthetic CO₂ fixation. Nat. Catal. 6, 1228–1240 (2023).
Article CAS MATH Google Scholar
Orsi, E., Claassens, N. J., Nikel, P. I. & Lindner, S. N. Growth-coupled selection of synthetic modules to accelerate cell factory development. Nat. Commun. 12, 5295 (2021).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Turlin, J., Dronsella, B., De Maria, A., Lindner, S. N. & Nikel, P. I. Integrated rational and evolutionary engineering of genome-reduced Pseudomonas putida strains promotes synthetic formate assimilation. Metab. Eng. 74, 191–205 (2022).
Article CAS PubMed Google Scholar
Delmas, V. A. et al. Genetic and biocatalytic basis of formate dependent growth of Escherichia coli strains evolved in continuous culture. Metab. Eng. 72, 200–214 (2022).
Article CAS PubMed MATH Google Scholar
Kim, K.-S. et al. The Rut pathway for pyrimidine degradation: novel chemistry and toxicity problems. J. Bacteriol. 192, 4089–4102 (2010).
Article CAS PubMed PubMed Central MATH Google Scholar
Pontrelli, S. et al. Metabolic repair through emergence of new pathways in Escherichia coli. Nat. Chem. Biol. 14, 1005–1009 (2018).
Article CAS PubMed MATH Google Scholar
Song, C. W., Kim, J. W., Cho, I. J. & Lee, S. Y. Metabolic engineering of Escherichia coli for the production of 3-hydroxypropionic acid and malonic acid through β-alanine route. ACS Synth. Biol. 5, 1256–1263 (2016).
Article CAS PubMed MATH Google Scholar
Keseler, I. M. et al. The EcoCyc database in 2021. Front Microbiol 12, 711077 (2021).
Article PubMed PubMed Central Google Scholar
Li, G.-W., Burkhardt, D., Gross, C. & Weissman, J. S. Quantifying absolute protein synthesis rates reveals principles underlying allocation of cellular resources. Cell 157, 624–635 (2014).
Article CAS PubMed PubMed Central Google Scholar
Schulz-Mirbach, H. et al. On the flexibility of the cellular amination network in E. coli. Elife 11, e77492 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Pauli, G. & Overath, P. ato Operon: a highly inducible system for acetoacetate and butyrate degradation in Escherichia coli. Eur. J. Biochem 29, 553–562 (1972).
Article CAS PubMed MATH Google Scholar
Jenkins, L. S. & Nunn, W. D. Genetic and molecular characterization of the genes involved in short-chain fatty acid degradation in Escherichia coli: the ato system. J. Bacteriol. 169, 42–52 (1987).
Article CAS PubMed PubMed Central MATH Google Scholar
Jenkins, L. S. & Nunn, W. D. Regulation of the ato operon by the atoC gene in Escherichia coli. J. Bacteriol. 169, 2096–2102 (1987).
Article CAS PubMed PubMed Central MATH Google Scholar
Marlière, P. & Mutzel, R. Method and device for selecting accelerated proliferation of living cells in suspension. U.S. Patent 6,686,194 B1 (2004).
Pezo, V. et al. A metabolic prototype for eliminating tryptophan from the genetic code. Sci. Rep. 3, 1359 (2013).
Article CAS PubMed PubMed Central MATH Google Scholar
Maki, H. & Sekiguchi, M. MutT protein specifically hydrolyses a potent mutagenic substrate for DNA synthesis. Nature 355, 273–275 (1992).
Article ADS CAS PubMed Google Scholar
Taddei, F. et al. Counteraction by MutT protein of transcriptional errors caused by oxidative damage. Science 278, 128–130 (1997).
Article CAS PubMed MATH Google Scholar
Shimokawa, H., Fujii, Y., Furuichi, M., Sekiguchi, M. & Nakabeppu, Y. Functional significance of conserved residues in the phosphohydrolase module of Escherichia coli MutT protein. Nucleic Acids Res 28, 3240–3249 (2000).
Article CAS PubMed PubMed Central Google Scholar
Metzgar, D. & Wills, C. Evidence for the adaptive evolution of mutation rates. Cell 101, 581–584 (2000).
Article CAS PubMed MATH Google Scholar
Rychel, K. et al. Laboratory evolution reveals transcriptional mechanisms underlying thermal adaptation of Escherichia coli. Preprint at https://doi.org/10.1101/2024.02.22.581624 (2024).
Phaneuf, P. V., Gosting, D., Palsson, B. O. & Feist, A. M. ALEdb 1.0: a database of mutations from adaptive laboratory evolution experimentation. Nucleic Acids Res. 47, D1164–D1171 (2019).
Article PubMed Google Scholar
Callens, M. et al. Hypermutator emergence in experimental Escherichia coli populations is stress-type dependent. Evol. Lett. 7, 252–261 (2023).
Article PubMed PubMed Central MATH Google Scholar
Bedbrook, J. R. & Ausubel, F. M. Recombination between bacterial plasmids leading to the formation of plasmid multimers. Cell 9, 707–716 (1976).
Article CAS PubMed MATH Google Scholar
Mazin, A. V., Timchenko, T. V., Saparbaev, M. K. & Mazina, O. M. Dimerization of plasmid DNA accelerates selection for antibiotic resistance. Mol. Microbiol 20, 101–108 (1996).
Article CAS PubMed Google Scholar
Crozat, E., Fournes, F., Cornet, F., Hallet, B. & Rousseau, P. Resolution of multimeric forms of circular plasmids and chromosomes. Microbiol. Spectr. 2, https://doi.org/10.1128/microbiolspec.plas-0025-2014 (2014).
Zelcbuch, L. et al. Spanning high-dimensional expression space using ribosome-binding site combinatorics. Nucleic Acids Res. 41, e98 (2013).
Article CAS PubMed PubMed Central Google Scholar
Orth, J. D., Thiele, I., Palsson & B, Ø. What is flux balance analysis? Nat. Biotechnol. 28, 245–248 (2010).
Article CAS PubMed PubMed Central Google Scholar
Fernie, A. R. & Bauwe, H. Wasteful, essential, evolutionary stepping stone? The multiple personalities of the photorespiratory pathway. Plant J. 102, 666–677 (2020).
Article CAS PubMed Google Scholar
Eisenhut, M., Roell, M.-S. & Weber, A. P. M. Mechanistic understanding of photorespiration paves the way to a new green revolution. N. Phytol. 223, 1762–1769 (2019).
Article Google Scholar
Walker, B. J., VanLoocke, A., Bernacchi, C. J. & Ort, D. R. The costs of photorespiration to food production now and in the future. Annu Rev. Plant Biol. 67, 107–129 (2016).
Article CAS PubMed Google Scholar
Trudeau, D. L. et al. Design and in vitro realization of carbon-conserving photorespiration. Proc. Natl Acad. Sci. USA 115, E11455–E11464 (2018).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Marchal, D. G. et al. Machine Learning-Supported Enzyme Engineering toward Improved CO₂-Fixation of Glycolyl-CoA Carboxylase. ACS Synth. Biol. 12, 3521–3530 (2023).
Article CAS PubMed PubMed Central MATH Google Scholar
Dronsella, B. et al. One-carbon fixation via the synthetic reductive glycine pathway exceeds yield of the Calvin cycle. Nat. Microbiol. 10, 646–653 (2025).
Gleizer, S. et al. Conversion of Escherichia coli to generate all biomass carbon from CO₂. Cell 179, 1255–1263.e12 (2019).
Article CAS PubMed PubMed Central MATH Google Scholar
Ben Nissan, R. et al. Autotrophic growth of Escherichia coli is achieved by a small number of genetic changes. Elife 12, RP88793 (2024).
Article PubMed PubMed Central MATH Google Scholar
Gassler, T., Baumschabl, M., Sallaberger, J., Egermeier, M. & Mattanovich, D. Adaptive laboratory evolution and reverse engineering enhances autotrophic growth in Pichia pastoris. Metab. Eng. 69, 112–121 (2022).
Article CAS PubMed MATH Google Scholar
Schulz-Mirbach, H., Dronsella, B., He, H. & Erb, T. J. Creating new-to-nature carbon fixation: a guide. Metab. Eng. 82, 12–28 (2024).
Article CAS PubMed Google Scholar
Bar-Even, A. Formate assimilation: the metabolic architecture of natural and synthetic pathways. Biochemistry 55, 3851–3863 (2016).
Article CAS PubMed MATH Google Scholar
Anthony, C. How half a century of research was required to understand bacterial growth on C1 and C2 compounds; the story of the serine cycle and the ethylmalonyl-CoA pathway. Sci. Prog. 94, 109–137 (2011).
Article CAS PubMed PubMed Central MATH Google Scholar
Nattermann, M. et al. Engineering a new-to-nature cascade for phosphate-dependent formate to formaldehyde conversion in vitro and in vivo. Nat. Commun. 14, 2682 (2023).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Siegel, J. B. et al. Computational protein design enables a novel one-carbon assimilation pathway. Proc. Natl Acad. Sci. 112, 3704–3709 (2015).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Chen, F. Y.-H., Jung, H.-W., Tsuei, C.-Y. & Liao, J. C. Converting Escherichia coli to a synthetic methylotroph growing solely on methanol. Cell 182, 933–946.e14 (2020).
Article CAS PubMed Google Scholar
Keller, P. et al. Generation of an Escherichia coli strain growing on methanol via the ribulose monophosphate cycle. Nat. Commun. 13, 5243 (2022).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Schann, K. & Wenk, S. The Serine Shunt enables formate conversion to formaldehyde in vivo. Sustainable Microbiol. https://doi.org/10.1093/sumbio/qvae024 (2024).
He, H., Höper, R., Dodenhöft, M., Marlière, P. & Bar-Even, A. An optimized methanol assimilation pathway relying on promiscuous formaldehyde-condensing aldolases in E. coli. Metab. Eng. 60, 1–13 (2020).
Article CAS PubMed Google Scholar
Wu, T. et al. Engineering a synthetic energy-efficient formaldehyde assimilation cycle in Escherichia coli. Nat. Commun. 14, 8490 (2023).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Nattermann, M. et al. Engineering a highly efficient carboligase for synthetic one-carbon metabolism. ACS Catal. 11, 5396–5404 (2021).
Article CAS PubMed PubMed Central MATH Google Scholar
Chou, A., Lee, S. H., Zhu, F., Clomburg, J. M. & Gonzalez, R. An orthogonal metabolic framework for one-carbon utilization. Nat. Metab. 3, 1385–1399 (2021).
Article CAS PubMed Google Scholar
Lee, S. H. et al. Identification of 2-Hydroxyacyl-CoA synthases with high acyloin condensation activity for orthogonal one-carbon bioconversion. ACS Catal. 13, 12007–12020 (2023).
Article CAS MATH Google Scholar
Claassens, N. J. et al. Phosphoglycolate salvage in a chemolithoautotroph using the Calvin cycle. Proc. Natl Acad. Sci. USA 117, 22452–22461 (2020).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Khmelenina, V. N., Colin Murrell, J., Smith, T. J. & Trotsenko, Y. A. Physiology and Biochemistry of the Aerobic Methanotrophs. in Aerobic Utilization of Hydrocarbons, Oils and Lipids (ed. Rojo, F.) 1–25 (Springer International Publishing, Cham, 2018).
Koo, C. W. & Rosenzweig, A. C. Biochemistry of aerobic biological methane oxidation. Chem. Soc. Rev. 50, 3424–3436 (2021).
Article CAS PubMed PubMed Central MATH Google Scholar
Henard, C. A. et al. Ribulose-1,5-bisphosphate carboxylase/oxygenase (RubisCO) is essential for growth of the methanotroph methylococcus capsulatus strain bath. Appl Environ. Microbiol 87, e0088121 (2021).
Article PubMed MATH Google Scholar
Bouzon, M. et al. A synthetic alternative to canonical one-carbon metabolism. ACS Synth. Biol. 6, 1520–1533 (2017).
Article CAS PubMed MATH Google Scholar
Robinson, W. G. & Coon, M. J. Synthesis of malonic semialdehyde, β-hydroxypropionate, and β-hydroxyisobutyrate. Methods in Enzymology 6, 549–553 (Academic Press, 1963).
Jonsson, S., Ricagno, S., Lindqvist, Y. & Richards, N. G. J. Kinetic and mechanistic characterization of the formyl-CoA transferase from Oxalobacter formigenes. J. Biol. Chem. 279, 36003–36012 (2004).
Article CAS PubMed Google Scholar
Burgener, S., Cortina, N. S. & Erb, T. J. Oxalyl-CoA decarboxylase enables nucleophilic one-carbon extension of aldehydes to chiral α-hydroxy acids. Angew. Chem. Int. Ed. 59, 5526–5530 (2020).
Article CAS Google Scholar
Kim, Y. et al. Revealing reaction intermediates in one-carbon elongation by thiamine diphosphate/CoA-dependent enzyme family. Commun. Chem. 7, 1–10 (2024).
Article CAS Google Scholar
Peter, D. M., Vögeli, B., Cortina, N. S. & Erb, T. J. A chemo-enzymatic road map to the synthesis of coa esters. Molecules 21, 517 (2016).
Article PubMed PubMed Central MATH Google Scholar
Vergne-Vaxelaire, C. et al. Nitrilase activity screening on structurally diverse substrates: providing biocatalytic tools for organic synthesis. Adv. Synth. Catal. 355, 1763–1779 (2013).
Article CAS MATH Google Scholar
Kroeger, J. K., Zarzycki, J. & Fuchs, G. A spectrophotometric assay for measuring acetyl-coenzyme A carboxylase. Anal. Biochem 411, 100–105 (2011).
Article CAS PubMed Google Scholar
Lu, Y., Yao, D. & Chen, C. 2-Hydrazinoquinoline as a derivatization agent for LC-MS-based metabolomic investigation of diabetic ketoacidosis. Metabolites 3, 993–1010 (2013).
Article PubMed PubMed Central MATH Google Scholar
Kabsch, W. XDS. Acta Crystallogr D. Biol. Crystallogr 66, 125–132 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Agirre, J. et al. The CCP4 suite: integrative software for macromolecular crystallography. Acta Crystallogr D. Struct. Biol. 79, 449–461 (2023).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Adams, P. D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D. Biol. Crystallogr 66, 213–221 (2010).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr D. Biol. Crystallogr 66, 486–501 (2010).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Waterhouse, A. et al. SWISS-MODEL: homology modelling of protein structures and complexes. Nucleic Acids Res. 46, W296–W303 (2018).
Article CAS PubMed PubMed Central MATH Google Scholar
Jensen, S. I., Lennen, R. M., Herrgård, M. J. & Nielsen, A. T. Seven gene deletions in seven days: Fast generation of Escherichia coli strains tolerant to acetate and osmotic stress. Sci. Rep. 5, 17874 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Datsenko, K. A. & Wanner, B. L. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl Acad. Sci. USA 97, 6640–6645 (2000).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Bassalo, M. C. et al. Rapid and efficient one-step metabolic pathway integration in E. coli. ACS Synth. Biol. 5, 561–568 (2016).
Article CAS PubMed MATH Google Scholar
Stukenberg, D. et al. The Marburg Collection: a golden gate DNA assembly framework for synthetic biology applications in Vibrio natriegens. ACS Synth. Biol. 10, 1904–1919 (2021).
Article CAS PubMed MATH Google Scholar
Shenoy, A. R. & Visweswariah, S. S. Site-directed mutagenesis using a single mutagenic oligonucleotide and DpnI digestion of template DNA. Anal. Biochem 319, 335–336 (2003).
Article CAS PubMed Google Scholar
Braatsch, S., Helmark, S., Kranz, H., Koebmann, B. & Jensen, P. R. Escherichia coli strains with promoter libraries constructed by Red/ET recombination pave the way for transcriptional fine-tuning. Biotechniques 45, 335–337 (2008).
Article CAS PubMed Google Scholar
Hay, R. & Bond, M. Kinetics of the decarboxylation of acetoacetic acid. Aust. J. Chem. 20, 1823–1828 (1967).
Article CAS MATH Google Scholar
Deatherage, D. E. & Barrick, J. E. Identification of mutations in laboratory-evolved microbes from next-generation sequencing data using breseq. Methods Mol. Biol. 1151, 165–188 (2014).
Article CAS PubMed PubMed Central MATH Google Scholar
Riley, M. et al. Escherichia coli K-12: a cooperatively developed annotation snapshot-2005. Nucleic Acids Res. 34, 1–9 (2006).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Zor, T. & Selinger, Z. Linearization of the Bradford protein assay increases its sensitivity: theoretical and experimental studies. Anal. Biochem 236, 302–308 (1996).
Article CAS PubMed MATH Google Scholar
Ebrahim, A., Lerman, J. A., Palsson, B. O. & Hyduke, D. R. COBRApy: COnstraints-based reconstruction and analysis for python. BMC Syst. Biol. 7, 74 (2013).
Article PubMed PubMed Central Google Scholar
Fu, X., Gregory, L. M., Weise, S. E. & Walker, B. J. Integrated flux and pool size analysis in plant central metabolism reveals unique roles of glycine and serine during photorespiration. Nat. Plants 9, 169–178 (2023).
Article CAS PubMed Google Scholar
Sharkey, T. D. Estimating the rate of photorespiration in leaves. Physiologia Plant. 73, 147–152 (1988).
Article CAS MATH Google Scholar
Ferguson, S. J. The ups and downs of P/O ratios: (and the question of non-integral coupling stoichiometries for oxidative phosphorylation and related processes). Trends Biochemical Sci. 11, 351–353 (1986).
Article CAS MATH Google Scholar
Hinkle, P. C. P/O ratios of mitochondrial oxidative phosphorylation. Biochimica et. Biophysica Acta (BBA) - Bioenerg. 1706, 1–11 (2005).
Article CAS MATH Google Scholar
Monk, J. M. et al. iML1515, a knowledgebase that computes Escherichia coli traits. Nat. Biotechnol. 35, 904–908 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Nielsen, J. & Keasling, J. D. Engineering cellular metabolism. Cell 164, 1185–1197 (2016).
Article CAS PubMed Google Scholar
Price, M. N. et al. Oxidative pathways of deoxyribose and deoxyribonate catabolism. mSystems 4, e00297–18 (2019).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

The authors are grateful to Beau Dronsella, Helena Schulz-Mirbach, Nico Claassens and Elad Noor for helpful discussions on the project and critical reading of the manuscript; to Peter Claus and Änne Michaelis for expert assistance with LC-MS measurements; to Hendrik Westedt for assistance with chemical synthesis; to Daniel Schindler and María Sánchez for helpful discussions on next-generation sequencing; to Frederik V. Schmidt and Christian Scholz for assistance with ICP-OES measurements. The authors acknowledge the assistance of the staff scientists of the beamlines P13 and P14 operated by EMBL Hamburg at the PETRA III storage ring (DESY, Hamburg, Germany). This work was supported by the Max Planck Society and the European Union H2020 Program (project GAIN4CROPS, grant agreement no. 862087). A.S. acknowledges funding from the International Max Planck Research School for Primary Metabolism and Plant Growth.

Funding

Open Access funding enabled and organized by Projekt DEAL.

Author information

These authors contributed equally: Ari Satanowski, Daniel G. Marchal.
Deceased: Arren Bar-Even.

Authors and Affiliations

Max Planck Institute for Terrestrial Microbiology, Karl-von-Frisch-Str. 10, Marburg, Germany
Ari Satanowski, Daniel G. Marchal, Hai He, Henrik M. Petri, Maren Nattermann, Simon Burgener, Nicole Paczia, Jan Zarzycki & Tobias J. Erb
Max Planck Institute of Molecular Plant Physiology, Am Mühlenberg 1, Potsdam, Germany
Ari Satanowski, Vittorio Rainaldi & Arren Bar-Even
Génomique Métabolique, Genoscope, Institut François Jacob, CEA, CNRS, Univ Evry, Université Paris-Saclay, Evry-Courcouronnes, France
Alain Perret, Jean-Louis Petit, Madeleine Bouzon, Volker Döring, Ivan Dubois & Virginie Pellouin
Molecular Systems Biology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Nijenborgh 7, Groningen, Netherlands
Edward N. Smith & Matthias Heinemann
Center for Synthetic Microbiology (SYNMIKRO), Karl-von-Frisch-Straße 14, Marburg, Germany
Tobias J. Erb

Authors

Ari Satanowski
View author publications
Search author on:PubMed Google Scholar
Daniel G. Marchal
View author publications
Search author on:PubMed Google Scholar
Alain Perret
View author publications
Search author on:PubMed Google Scholar
Jean-Louis Petit
View author publications
Search author on:PubMed Google Scholar
Madeleine Bouzon
View author publications
Search author on:PubMed Google Scholar
Volker Döring
View author publications
Search author on:PubMed Google Scholar
Ivan Dubois
View author publications
Search author on:PubMed Google Scholar
Hai He
View author publications
Search author on:PubMed Google Scholar
Edward N. Smith
View author publications
Search author on:PubMed Google Scholar
Virginie Pellouin
View author publications
Search author on:PubMed Google Scholar
Henrik M. Petri
View author publications
Search author on:PubMed Google Scholar
Vittorio Rainaldi
View author publications
Search author on:PubMed Google Scholar
Maren Nattermann
View author publications
Search author on:PubMed Google Scholar
Simon Burgener
View author publications
Search author on:PubMed Google Scholar
Nicole Paczia
View author publications
Search author on:PubMed Google Scholar
Jan Zarzycki
View author publications
Search author on:PubMed Google Scholar
Matthias Heinemann
View author publications
Search author on:PubMed Google Scholar
Arren Bar-Even
View author publications
Search author on:PubMed Google Scholar
Tobias J. Erb
View author publications
Search author on:PubMed Google Scholar

Contributions

A.B.-E. and A.S. conceived the study; A.B-E. and A.S. designed and analyzed CO₂-reduction cycles with contributions from S.B.; A.P., J-L.P. and V.P. designed and performed the initial in vitro screen of the BKACE collection; M.N. performed in vitro experiments on the reductive formamide pathway; D.G.M. designed and performed all other in vitro experiments, optimized the BKACE screening assay, synthesized and purified CoA-esters, MSA and 1,3-¹³C₂-acetoacetate, purified enzymes and crystallized BKACE; D.G.M. and J.Z. collected and processed x-ray diffraction data, refined and analyzed crystal structures; A.S. designed and performed all in vivo experiments, designed ALE experiments, and analyzed genome sequencing data; M.B., V.D. and I.D. performed turbidostat evolutions; H.H., V.R. and H.P. contributed to strain engineering and in vivo experiments; A.S. and N.P. performed isotope tracing experiments and LC-MS analysis; H.H., E.S. and A.S. performed stoichiometric modeling; T.J.E., A.B-E. and M.H. supervised research and acquired funding; A.S., D.G.M. and T.J.E wrote the manuscript with contributions from all authors.

Corresponding authors

Correspondence to Ari Satanowski or Tobias J. Erb.

Ethics declarations

Competing interests

T.J.E. and A.B-E. are named inventors on the patent US10781456B2, describing photorespiration bypasses. The other authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer review file

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Supplementary Data 5

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Satanowski, A., Marchal, D.G., Perret, A. et al. Design and implementation of aerobic and ambient CO₂-reduction as an entry-point for enhanced carbon fixation. Nat Commun 16, 3134 (2025). https://doi.org/10.1038/s41467-025-57549-4

Download citation

Received: 03 June 2024
Accepted: 25 February 2025
Published: 01 April 2025
DOI: https://doi.org/10.1038/s41467-025-57549-4

This article is cited by

A blueprint for designing the next-generation of synthetic C1 microbes
- Giusi Favoino
- Òscar Puiggené
- Pablo I. Nikel
Nature Communications (2025)
Planktonic anammox bacteria toward a better understanding of ecophysiological traits and harnessing the untapped potential as a bioresource
- Yantong Liu
- Mingda Zhou
- Yayi Wang
Bioprocess and Biosystems Engineering (2025)