The fed-batch bioprocess for leucine production with M. marburgensis<\/em> mutant strains in bioreactors has proven to be scalable to 150 L pilot-plant scale without substantial losses in leucine quality and quantity. Feeding schemes for trace metals, nitrogen and sulfur sources, and H2 and CO2 need to be investigated further to enhance volumetric leucine productivity and specific productivity to prove the economic feasibility of the process. Additionally, validation of fermentation campaigns, microbial strain stability, and further scaling of the bioprocess are essential next steps. Therefore, an overall TRL of 5 for the leucine production bioprocess is estimated. The downstream processing for leucine purification could, in principle, be adopted from established glucose-based leucine production platforms once product concentrations reach a level of 10 g L\u22121 or higher. Therefore, the downstream processing of leucine from fermentation broth ranges between TRL 8 and TRL 9.<\/p>\n\n\n\n Regarding the integration of the bioprocess with steam methane reformer, a TRL between TRL 8 and TRL 9 for the steam methane reformer process is assumed. In terms of regulatory requirements for full-scale implementation of the presented process, initial insights suggest the feasibility of GRAS approval for M. marburgensis<\/em>.<\/p>\n\n\n\n Beyond their nutritional value, amino acids are of relevance in medicine and, due to their chemical properties, they are indispensable for applications in many different realms, such as pharmaceutics, cosmetics, animal feed, food, and the beverage industry. In this research article, we report an archaeal cell factory for leucine production from CO2 that has been generated by rational design, random mutagenesis, and pathway engineering. The cell factory has been bioprocess-technologically examined and successfully scaled up with regard to productivity, product quality, and operational stability. In a 2-day fed-batch campaign, we produced 181 g of leucine from CO2 at the 150-L pilot-plant scale with a mean volumetric leucine productivity of 65 mg L\u22121 h\u22121. A thorough techno-economic analysis indicates that the rollout of leucine production from CO2 is nearly economically feasible on a global scale.<\/p>\n\n\n\n Twenty proteinogenic amino acids (AAs) are relevant for feedstock and human nutrition. Additionally, AAs are of importance in several industrial sectors [1,2]. In particular, the essential AAs, including the branched-chain amino acids (BCAAs) (valine, leucine, and isoleucine), which cannot be synthesized by animals or humans, are of interest for industrial AA production. Beyond their essential role, BCAAs are valuable in sports nutrition for enhancing muscle growth, reducing fatigue, and accelerating recovery after exercise [3\u20135]. Additionally, there is evidence that dietary leucine might be associated with fat reduction [6,7]. The large variety of applications for AAs has resulted in a global market size of USD 32.19 billion in 2025, which is projected to grow to USD 56.61 billion by 2032, with a robust compound annual growth rate (CAGR) of 8.4%(https:\/\/www.precedenceresearch.com\/amino-acids-market). Leucine, the key BCAA relevant for this study, has a market size of USD 1.23 billion in 2025 and is expected to surge to USD 3.67 billion by 2032, reflecting an impressive CAGR of 16.89% (https:\/\/www.marketreportanalytics.com\/reports\/leucine-262237, https:\/\/www.researchnester.com\/reports\/leucine-market\/2847).<\/p>\n\n\n\n Nowadays, most of the essential AAs, including leucine, are produced in microbial fermentation campaigns [8\u201311]. The main microbial cell factories for AA production are Escherichia coli and Corynebacterium glutamicum. In these species, genetically engineered strains have been optimized over decades for glucose-based production of individual AAs [11\u201313]. However, due to foreseeable climate-crisis-based restrictions in agricultural land use for glucose production [14,15], valorization of alternative carbon sources and C1 compounds, such as hemicellulose, methanol, or carbon dioxide (CO2), is being considered for next-generation AA production. To date, approaches for the use of alternative carbon sources, such as xylose or arabinose, in existing E. coli and C. glutamicum cell factories for AA production have been considered [16,17]. Additionally, C1-fixing microbes as hosts for AA production, such as Bacillus methanolicus for lysine production from methanol, or Cupriavidus necator and Synechocystis spp. for valine and aromatic AA production from CO2, are employed [9,18\u201320]. Recently, however, it has been found that CO2-fixing methanogenic archaea (methanogens) naturally secrete AAs [21\u201323].<\/p>\n\n\n\n One of them is Methanothermobacter marburgensis (see Glossary), a thermophilic methanogen that has already attracted attention as a microbial cell factory for gas-fermentative methane production in Power-to-Gas (P2G) systems [24\u201326]. In this regard, molecular hydrogen (H2) generated from renewable energy sources (\u2018green H2\u2019), such as wind or solar energy, is used to reduce carbon capture-derived or waste-gas-derived CO2. Hence, H2 is biologically converted to methane in P2G, or, in Power-to-Chemicals, to, for example, AAs [27,28]. Within the P2G and AA production frameworks, it has already been proven that bioprocesses with M. marburgensis are scalable [29]. Moreover, M. marburgensis is highly resistant to shear forces [24] due to the rigid characteristics of the archaeal peptidoglycan-harboring cell wall [30]. Recently, markerless mutagenesis tools have been developed that allow metabolic and pathway engineering of M. marburgensis to improve production rates of individual AAs [31,32]. Taken together, M. marburgensis is a highly suitable chassis for archaeal cell factory development for AA production from CO2.<\/p>\n\n\n\n As a first proof of concept, we chose the production of the BCAA leucine due to its relevance in human sports nutrition and its application as an additive in animal feed [33]. On a physiological level, we aimed to increase the carbon flux from CO2 to leucine. In a two-step metabolic engineering attempt, an archaeal cell factory was developed through rational design and random mutagenesis. On a bioprocess-technological level, the leucine cell factory has been proven in various gas fermentation campaigns regarding bioprocess stability and functionality for leucine production in lab-scale continuous culture and fed-batch in various bioreactors up to the 150 L pilot-plant scale. Through a techno-economic model (TEM), we put the leucine productivity and leucine concentrations from the pilot scale into an economic context to assess the economic feasibility and the putative commercial impact of our technology on a global scale.<\/p>\n\n\n\n In the first step, we evaluated whether leucine overproduction and secretion by M. marburgensis<\/em> could be achieved using a random mutagenesis approach to increase carbon flux toward leucine. Through high-throughput screening of mutagenized strains, we found significantly higher leucine concentrations, with up to eightfold and 15-fold increases (P < 0.01) in the specific leucine production rate (qLeu) and selectivity, respectively, compared with wild-type M. marburgensis (Figures S1\u2013S3 in the Supplemental information online). All other detected AAs and the strains\u2019 growth behavior were not significantly impacted by the overproduction of leucine (Figures S1\u2013S3). Sequencing of three respective overproduction mutants revealed single-nucleotide polymorphisms (SNPs) in each of their isopropylmalate synthase (IPMS)-encoding genes at a position relevant for release of an allosteric leucine inhibition (L461F and A430V), but no SNPs in the acetohydroxyacid synthase (AHAS)-encoding genes were detected (Figure S4 in the Supplemental information online) (Figure 1D).<\/p>\n\n\n Following the initial proof of concept for leucine overproduction, we tested whether further increased carbon flux to valine and leucine was achievable by rational design, specifically by genetic engineering via overexpression of the native AHAS0 from Methanothermobacter thermautotrophicus<\/em> in M. marburgensis<\/em> (Figure 1D). Here, we demonstrated significantly increased specific valine production rates (qVal) and qLeu production rates increased by 29- and 31-fold (P < 0.01, n = 3), respectively, with the M. marburgensis<\/em> AHAS0 mutant compared with the wild-type strain (Figure 1E,F). To analyze the impact of allosteric leucine or valine inhibition on the heterologous AHAS0, we performed an in vitro enzyme assay and found 50% and 35% of inhibition, respectively (Figure 1B,C). Thus, although AHAS0 was partially inhibited, these findings render the enzyme still reasonably functional at high leucine and valine concentrations.<\/p>\n\n\n\n To combine the effects of random mutagenesis and overexpression of AHAS0 to enhance leucine production, we performed random mutagenesis on M. marburgensis<\/em> AHAS0 (Figure 1A). After screening 36 mutants, we characterized the three most promising leucine-producing mutants based on leucine specificity and concentration in closed batch experiments (n = 3) (Figures S5\u2013S7 in the Supplemental information online). We found a two- to threefold increase in qLeu and a decrease in qVal of up to fivefold (Figure 1E,F) compared with M. marburgensis<\/em> AHAS0. That marks a selectivity toward leucine compared with all other 20 AAs of up to 67% (Figures S5\u2013S7). Interestingly, we found SNPs leading to G458S, G426R, and A461V in the ipms gene of the randomly mutagenized M. marburgensis<\/em> AHAS0 mutants E3, E4, and A7, respectively. These SNPs occurred in similar locations as they had been already confirmed in allosteric released IPMS from E. coli <\/em>or C. glutamicum<\/em> (Figure 1H) [34\u201336]. To support our findings, we confirmed allosteric leucine binding sites in IPMS with docking assays using AutoDock Vina. <\/p>\n\n\n\n Although the regulatory domain of IPMS is less conserved across the bacterial and archaeal domains compared with the catalytic domain of IPMS, the allosteric binding site of leucine appears to be highly similar (Figure 1G). In addition, we performed enzyme assays for leucine inhibition of the M. marburgensis AHAS0 mutant A7 IPMS and wild-type M. marburgensis IPMS. The results indicate substantial reduction in the leucine allosteric product inhibition in the truncated enzyme variant activity compared with the wild-type IPMS (Figure S8 in the Supplemental information online). Hence, the randomly mutagenized M. marburgensis AHAS0 mutants E3, E4, and A7 were considered the first archaeal leucine cell factory and transferred bioreactor systems for continuous cultivation.<\/p>\n\n\n\n …<\/p>\n\n\n\nHighlights<\/h3>\n\n\n\n
\n
By combining rational design and random mutagenesis, we establish the first archaeal cell factory for leucine production.<\/li>\n\n\n\nAbstract<\/h3>\n\n\n\n
Graphical Abstract<\/h3>\n\n\n\n
![\"\"](\"https:\/\/renewable-carbon.eu\/news\/media\/2026\/08\/ga1.jpg\")
Introduction<\/h3>\n\n\n\n
Results<\/h3>\n\n\n\n
A combination of genetic engineering and random mutagenesis leads to leucine overproduction in M. marburgensis<\/em> <\/h3>\n\n\n\n
![\"\"](\"https:\/\/renewable-carbon.eu\/news\/media\/2026\/08\/Bildschirmfoto-2026-08-19-um-10.44.40-1024x966.png\")