Engineering Artificial Metabolons via Phase-Separated Proteins for Metabolic Regulation in Yarrowia lipolytica
Abstract:
Metabolic engineering aims to develop efficient microbial cell factories to produce value-added chemicals. However, pathway efficiency is frequently limited by the free diffusion of enzymes, byproduct accumulation and competing pathways. Organizing pathway enzymes into artificial metabolons via liquid-liquid phase separation (LLPS) offers a route to overcome these diffusion barriers by creating membrane-less compartments. In this study, we explored the construction of a synthetic multi-enzyme condensate platform in the oleaginous yeast Yarrowia lipolytica using the known phase-separating protein hnRNPA1.
In vivo confocal imaging showed that hnRNPA1 forms visible droplets, and expressing target small RNA (sRNA) significantly enhanced droplet assembly. Fusing hnRNPA1 drove normally diffuse cytosolic lipogenic enzymes (acetyl-CoA carboxylase ACC1 and fatty acid synthase FAS1) into shared condensates. To test substrate channeling without interference from high native lipid levels, the triacetic acid lactone (TAL) pathway catalyzed by 2-pyrone synthase (2PS) was used as a reporter. However, multi-enzyme fusions with hnRNPA1 caused an unexpected drop in TAL titers relative to the free enzymes, with sRNA-free strains yielding higher TAL-to-fatty acid ratios than those with sRNA.
Fluorescence recovery after photobleaching (FRAP) revealed a mobile fraction of only ~5% (95% immobile) in sRNA-induced condensates, indicating that the droplets transitioned into a solid-like gel state, which shifted catalysis into a mass-transfer-limited regime. This gelation restricted substrate uptake or product release inside the condensates, while the bulky hnRNPA1 tags blocked active sites via steric hindrance.
These results reveal key physical limitations of protein-based phase separation, showing that maintaining droplet fluidity and minimizing steric hindrance are critical when building functional synthetic metabolons. Future efforts will optimize the hnRNPA1 sequence and use peptide tags to recruit enzymes to the condensate surface.

