Research
Systems Biology:
Control of cellular metabolism
Metabolism is staggeringly complex. Hundreds of biochemical reactions run inside a single cell at the same time, many of them at very high rates, each one feeding the next. How a bacterium like E. coli keeps this from descending into chaos is still an open question. Imagine that just one enzyme does not work properly. In principle this would block a pathway, metabolites would accumulate, and the whole network would fail. Yet we know this almost never happens. Cellular metabolism is robust and maintains it function in the face of internal and external perturbations. But how?
The picture shows amino acid pathways in E. coli and their regulation. We asked why these pathways have so many feedbacks. Turns out they solve a trade-off between efficiency and robustness (from Timurs Paper in Cell Systems, Sander et al 2019).
To find this out, we used CRISPR interference to knock down essentially every metabolic enzyme, one at a time, and then study the consequences across the whole network with metabolomics and proteomics (Donati et al 2021, Rapp et al 2026). This systems-level view revealed a set of design principles for how cells stay stable and efficient. Metabolic networks are remarkably robust: they buffer large decreases in enzyme levels rather than collapsing. Enzymes are held in deliberate overabundance, so there is always overcapacity when conditions change or gene expression is noisy (Sander et al 2019). Moreover, cells deliberately keep most metabolites at low concentrations, because high levels trigger regulatory crosstalk and unwanted side reactions (Rapp et al 2026).
Much of this robustness comes from a constant interplay between metabolism and the genome: metabolites act as signals that switch genes on and off, while gene expression sets how much of each enzyme a cell makes. To map this two-way control, we developed methods that identify which metabolites regulate which transcription factors, predicting new metabolite–regulator interactions and capturing known ones from data alone (Lempp et al 2020).
Such large scale perturbation data will transform both mechanistic metabolic modelling approaches and machine learning methods.
Synthetic Biology: Switching bacteria between
growth and production of chemicals
Industrial microbes face a fundamental conflict: growth and overproduction are not compatible. Growth drains resources from production. And overproduction leads to instabilities and genetic drifts. Our synthetic biology program resolves this with so-called two-stage processes (Shabestary et al 2024). First, we grow cells to high densities without the pressure of producing a chemical. Next, we switch them into a non-growing state where metabolism maintains active and the product is made with high rates.
The picture shows our mini-bioreactor system that automatically induces production and growth arrest by switching the temperature when when a specific biomass concentration is reached.
The SynBio parts that make this possible are our thermo-switches: enzymes that are engineered so that a simple change in temperature turns a cellular process on or off (Schramm et al 2023). We build them at genome scale. Starting from a CRISPR library of over 15,000 E. coli mutants, each carrying a single amino-acid change in an essential protein, we identified more than a thousand temperature-sensitive variants and turned them into thermo-switches. We build switches that work in both directions off-switches that shut a process down, and on-switches that activate a process. All based on a simple signal: temperature. This gives us control over virtually every function in the cell and opens up new ways for control of bioprocesses.
We have used this control to keep cells metabolically active and produce valuable compounds such as the amino acids arginine and citrulline (Schramm et al 2020, Schramm et al 2023). Along the way we learned that successful engineering is about balance, not brute force overexpression: partially knocking genes down with CRISPR interference (Sander et al 2019), or relieving the hidden metabolic burden of synthetic pathways (Wang et al 2019), often outperforms simply deleting or overexpressing genes.
We are now combining these switches with sustainable feedstocks and move from glucose towards CO₂, formate, and ethanol, to build bacteria that make chemicals and bioplastics for a sustainable, low-carbon bioeconomy.
Metabolic Engineering:
Sustainable production of bioplastics and amino acids from CO2
Almost all industrial fermentation today runs on sugar from crops. But sugar competes with food, needs farmland, and has a heavy carbon footprint (e.g. due to fertilizers). We are engineering E. coli to run instead on the simplest and most sustainable carbon sources, such as CO2 itself, and one and two-carbon molecules formate (C1) and ethanol (C2). If our bacteria can grow and produce on these feedstocks, bioprocesses become not just carbon-neutral but potentially carbon-negative. With the lab of Ron Milo at the Weizmann Institute, we are turning autotrophic E. coli strains that grow on CO2 alone into cell factories for biobased production of chemicals (Nissan et al 2024). And we are also building E. coli strains that consume ethanol and convert it into products. A main focus is on production of amino acids, but in a new project we make the biodegradable bioplastic cyanophycin, which is a polymer of the amino acids aspartate and arginine. The goal is to produce it from ethanol using our temperature-controlled two-stage processes to switch cells from growth into production. To achieve all this we use our CRISPR and metabolomics methods to find and fix the bottlenecks in CO₂ fixation and feedstock use.
Infection Biology: Antibiotic resistance and host-pathogen interactions
Antimicrobial resistance is one of the largest threats to global health, and tackling it requires understanding bacteria not in isolation but as pathogens that interact with antibiotics, with each other, and with their host. We study how bacterial pathogens respond to antibiotic treatment, how resistance arises, and how the metabolic state influences antibiotic action.
Working in E. coli and Staphylococcus aureus, we combine genomics, proteomics, and metabolomics to follow what actually happens inside bacteria when they are challenged with antibiotics. This lets us watch resistance evolve in real time for example, how surviving subpopulations regrow after treatment and acquire the mutations that make resistance permanent and to identify the cellular states that allow some cells to tolerate antibiotics that usually kill them (Verhuelsdonk et al 2026). The cells own metabolism is one of several factors that can tip this balance, alongside classical resistance mechanisms and stress responses (Lubrano et al 2025).
Much of this work is collaborative, embedding our measurement expertise within the wider infection-biology community in Tübingen. Together with partners we study how signalling molecules govern antibiotic tolerance, how commensal and pathogenic bacteria coexist, and how an ecological view of the microbial community can inform new strategies against infection (Maier et al 2024).
Our goal is to understand the mechanisms of antibiotic action and resistance well enough to make existing drugs work better and last longer and to help anticipate, rather than chase, the next wave of resistance.