These are the *minimal* fluxes if flux distributions would be optimal. In reality, flux through BKACE might be higher, e.g. due to losses/inefficiencies/imbalances down the line. Also worth noting: It's "only" 8% because formate supplies "only" C1-metabolism and serine. Happy to discuss more.

Great question, Noam! Depends a bit on the cycle segment, and how we define "% of biomass flux". In our final implementation (acetoacetate as main carbon source), "module 3" of the cycle carries 100% biomass flux. The other modules carry 8% (when normalized to acetoacetate influx). FBA estimation:

📖 Read the full story here, and let us know what you think:
nature.com/articles/s41...

I’m deeply grateful to my PhD advisors Tobi Erb and Arren Bar-Even, who enabled this work and helped shaping the team around it.

Shout-out to Daniel Marchal (not on Bluesky yet 😉), who led the study together with me. And huge thanks to our many collaborators – this was my first experience at bringing together so many people with different expertise for a shared project.

Where do we go from here?
I hope our work shows that aerobic CO2-reduction is a metabolic route worth considering for various bioengineering goals: From boosting plant yields 🌱🌻 (by enabling an improved photosynthesis or photorespiration), to creating entirely new modes of synthetic autotrophy 💨🧬🦠

Eventually, we assembled the complete CORE cycle, successfully supplying E.coli's C1-metabolism (nucleotides etc.) and serine biosynthesis from CO2. It ain't a new autotrophic E.coli just yet, but we'll get there...⏳📆

To show that our metabolic designs can make it from the drawing board to an actual living cell, we used growth-coupled selection and adaptive evolution in E.coli.
Breaking the cyclic pathway down into short, testable “modules” was key for this strategy. 🧩🔄🗝️

In the paper, we describe how we approached this: step-by-step from in silico enzyme selection 🖥️🧠, to in vitro assays 🧪📈, to testing in living cells 🦠🧬.
For our most promising pathway design, we screened more than 120 enzyme homologues to find the desired enzyme candidates.

To get to this point, we drafted a dozen hypothetical pathways that we envisioned would allow CO2 reduction. Most of them required occasional “dream reactions” 😶‍🌫️🧠💭 – novel enzymatic activities not known to exist, yet. The tough part was how to find and realize those! 🔎🕵️‍♀️

We’ve now unlocked this ability for fully aerobic organisms and ambient CO2 levels. 🫁 🌍
The newly designed “CORE cycle” converts CO2 to formate without the need for any oxygen-sensitive enzymes or low-redox potential electron donors. Best of all: We’ve already made it work in E. coli.

But: Such “reduction first” CO2 fixation pathways have so far remained limited to oxygen-free (anaerobic) and high-CO2 environments – Everywhere else, enzymatic reduction of CO2 to formate is a tough uphill battle (thermodynamically speaking), especially in living cells.

What’s the secret to energy-efficient biological CO2 assimilation?
Turns out, anaerobic organisms have known for billions of years: Enzymatically convert CO2 to formate first, before upgrading it to biomass – all within the same cell! That gives rise to unrivalled ATP-efficiency.