CO2 fixation gets a second chance

Synthetic metabolic pathways that circumvent photorespiration can improve crop growth. Now, an efficient photorespiration bypass with a new-to-nature carboxylation step has been engineered and demonstrated in vitro.

Retrobiosynthesis is a powerful tool in metabolic engineering, where it is used to design new metabolic pathways for production or consumption of target compounds. The concept of retrobiosynthesis is beguilingly simple: a target is envisioned and potential steps for its production (or consumption) are enumerated by algorithms trained on known biochemical transformations1. In practice, retrobiosynthesis pathways often suggest transformations that have not been reported previously, or with no known enzyme. Selecting and engineering enzymes for such non-natural reactions poses a major challenge to implementation.

One emerging area for retrobiosynthesis is in approaches to bypass photorespiration, the series of reactions in plants that work to metabolize RuBisCO’s oxygenation product, 2-phosphoglycolate (2-PG). Through photorespiration, two molecules of 2-PG are salvaged to synthesize RuBisCO’s carboxylation product, the triosephosphate 3-phosphoglycerate (3-PG). Photorespiration is ATP-intensive and involves loss of nitrogen and previously fixed CO2, resulting in an estimated 20–50% reduction in photosynthetic efficiency in C3 plants such as wheat. Early synthetic 2-PG salvage pathways, when transformed into tobacco, boosted growth up to 24% in field trials2. Such successes may be just the beginning. A few proposed 2-PG salvage pathways include carboxylation steps, which make them carbon-positive3,4. Among these, the TaCo pathway, named for its tartronyl-CoA intermediate, is particularly attractive because it would enable an efficient five-step transformation of 2-PG to 3-PG (Fig. 1)4. However, two of the enzymatic steps in the TaCo pathway were not known to exist in nature.

Fig. 1: Realizing the tartronyl-CoA pathway.

a, The tartronyl-CoA (TaCo) pathway is designed to convert RuBisCO’s oxygenation product, 2-phosphoglycolate, into 3-phosphoglycerate with a novel carbon fixation step. Suitable GCS and TCR activities were found by screening for promiscuous activity among reported acetyl-CoA synthetases and malonyl-CoA reductases, respectively. The carboxylation of glycolyl-CoA by GCC has not been previously reported for any enzyme. b, Retrobiosynthetic pathways such as TaCo are realized by following a cycle of design, build, and test. c, The starting point for engineering a GCC was the biotin- and ATP-dependent propionyl-CoA carboxylase of Methylorubrum extorquens (MePCC). A cryo-EM structure of MePCC facilitated three point mutations to accommodate glycolyl-CoA, resulting in GCC M3. Two rounds of error-prone PCR mutagenesis of GCC M3 revealed GCC M5, a five-point mutant with higher activity and less futile ATP hydrolysis. RuBP, ribulose 1,5-bisphosphate; GCS, glycolyl-CoA synthetase; GCC, glycolyl-CoA carboxylase; TCR, tartronyl-CoA reductase; Pi, inorganic phosphate; PPi inorganic pyrophosphate.

Now, writing in Nature Catalysis, Erb and colleagues have realized the TaCo pathway in vitro, through state-of-the-art enzyme engineering and directed evolution5.

At the heart of the TaCo pathway is the carboxylation of glycolyl-CoA with bicarbonate, forming tartronyl-CoA, a compound not known to occur in nature. As a starting point for crafting a glycolyl-CoA carboxylase (GCC), the authors turned to propionyl-CoA carboxylase (PCC), an ATP- and biotin-dependent carboxylase central to organic acid assimilation in some bacteria. Based on structural similarity to propionyl-CoA, carboxylation of the smaller glycolyl-CoA was expected to occur with similar chemistry. Of four screened PCCs, only PCC from Methylorubrum extorquens (MePCC) showed detectable activity toward glycolyl-CoA. However, even low activity is sufficient to begin an enzyme engineering campaign. The authors collected cryogenic electron microscopy (cryo-EM) data on the MePCC complex to guide rational mutation of the enzyme’s carboxytransferase subunit. Three mutations were performed to accommodate and position glycolyl-CoA and to hydrogen-bond a hydroxyl group not present in propionyl-CoA. These mutations improved catalytic efficiency 50-fold. Next, error-prone PCR (epPCR) on the carboxytransferase subunit was performed. As is typical with epPCR, beneficial mutations were expected to be sparse, and there was no obvious high-throughput screen for GCC activity that would allow isolation of improved variants. To estimate the positive hit rate in the GCC mutant library, the authors devised a droplet-microfluidics assay linked to tartronyl-CoA production, and screened single E. coli cells harbouring GCC mutants. The assay was designed to be ATP-limited so as to select GCC variants with more efficient ATP usage. Of 100,000 screened mutants, 4.7% gave more product than the parent, sufficient to give confidence that more ATP-efficient mutants could also be isolated through colony screening. After two rounds of directed evolution, a GCC with two additional mutations was isolated with a further 10-fold higher catalytic efficiency (kcat/KM = 3.6 × 104 M–1 s–1), similar to natural acyl-CoA carboxylases.

The evolved GCC was then shown, through isotope-label tracing, to function in concert with the other enzymes of the TaCo pathway to convert 2-PG to glycerate in vitro. Finally, the authors connected the TaCo pathway to their recently developed synthetic CO2-to-glycolate crotonyl-coenzyme A (CoA)/ethylmalonyl-CoA/hydroxybutyryl-CoA (CETCH) cycle6, resulting in a 21-enzyme mixture that catalysed in vitro conversion of CO2 to glycerate, powered by ATP derived from added phosphocreatine. In the first iteration, promiscuity of a CETCH enzyme with TaCo intermediates (and vice versa) led to dead-end metabolites, which reduced yields. After identification of this intermediate with LC-MS, the troublesome enzyme was replaced by a homologue with less promiscuity; the successful debugging of the CETCH-TaCo cocktail is a testament to analytical skill.

Engineering acyl-CoA carboxylases is not easy. The bacterial enzyme assembles as a three-protein, multimeric complex and catalyses a two-step, double displacement reaction. Biotin is first carboxylated in an ATP-dependent mechanism, and the resulting carboxybiotin is shuttled to the acyl-CoA substrate. Furthermore, previous kinetic studies suggest that the enzyme uses biotin-induced conformational gating to prevent futile ATP hydrolysis7. In the present study, the authors found that futile ATP hydrolysis by MePCC was high when acting on the non-native substrate glycolyl-CoA, about 100 ATPs were hydrolysed per carboxylated substrate. Interestingly, in the most active GCC mutant, ATP usage was reduced to 5 ATPs per substrate, even though the biotin carboxylase subunit was not targeted for mutation. More research is needed on the conformational gating that prevents futile ATP hydrolysis in acyl-CoA carboxylases, and the cryo-EM-derived models of MePCC and the evolved GCC presented here may help. In the meantime, futile ATP hydrolysis complicates using ATP demand as a criterion for comparing synthetic pathways.

The report by Erb and colleagues presents a blueprint for how to design and achieve a synthetic metabolic pathway containing new reactions. The next step is to test the TaCo pathway in vivo, which will bring both benefits and challenges, in particular cross-reactivity with native metabolism. In bacteria, TaCo pathway activity could be connected to cell growth; for example, through insertion into a 3-phosphoglycerate auxotroph, which could put selection pressure on more active variants. In plants, the enzymes will likely be targeted to the chloroplast and optimization of enzyme levels, as well as substrate and co-enzyme supply, can be anticipated. Because native photorespiration in C3 plants is intertwined with amino acid biosynthesis, it is not yet clear whether a synthetic bypass to salvage glycolate can fully replace it2. Perturbation of the Calvin cycle must be considered carefully, as cycle stability can be negatively affected by a forced alteration of metabolite levels8. The TaCo pathway can alleviate a major constraint in RuBisCO-driven carbon fixation. In light of its potential benefit for crop yields, and with in vitro integration now demonstrated, it is poised to become a crowning achievement for retrobiosynthesis.


  1. 1.

    Hafner, J., MohammadiPeyhani, H., Sveshnikova, A., Scheidegger, A. & Hatzimanikatis, V. ACS Syn. Biol. 9, 1479–1482 (2020).

    CAS  Article  Google Scholar 

  2. 2.

    South, P. F., Cavanagh, A. P., Liu, H. W. & Ort, D. R. Science 363, eaat9077 (2019).

    CAS  Article  Google Scholar 

  3. 3.

    Shih, P. M., Zarzycki, J., Niyogi, K. K. & Kerfeld, C. A. J. Biol. Chem. 289, 9493–9500 (2014).

    CAS  Article  Google Scholar 

  4. 4.

    Trudeau, D. L. et al. Proc. Natl Acad. Sci. USA 115, E11455–E11464 (2018).

    CAS  Article  Google Scholar 

  5. 5.

    Scheffen et al. Nat. Catal. (2020).

  6. 6.

    Schwander, T., Schada von Borzyskowski, L., Burgener, S., Cortina, N. S. & Erb, T. J. Science 354, 900–904 (2016).

    CAS  Article  Google Scholar 

  7. 7.

    Waldrop, G. L., Holden, H. M. & St Maurice, M. Protein Sci. 21, 1597–1619 (2012).

    CAS  Article  Google Scholar 

  8. 8.

    Janasch, M., Asplund-Samuelsson, J., Steuer, R. & Hudson, E. P. J. Exp. Bot. 70, 973–983 (2019).

    CAS  PubMed  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Elton P. Hudson.

Ethics declarations

Competing interests

The authors declare no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Janasch, M., Hudson, E.P. CO2 fixation gets a second chance. Nat Catal 4, 94–95 (2021).

Download citation


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing