No abstract

Through advanced mechanistic modeling and the generation of large high-quality datasets, machine learning is becoming an integral part of understanding and engineering living systems. Here we show that mechanistic and machine learning models can be combined to enable accurate genotype-to-phenotype predictions. We use a genome-scale model to pinpoint engineering targets, efficient library construction of metabolic pathway designs, and high-throughput biosensor-enabled screening for training diverse machine learning algorithms. From a single data-generation cycle, this enables successful forward engineering of complex aromatic amino acid metabolism in yeast, with the best machine learning-guided design recommendations improving tryptophan titer and productivity by up to 74 and 43%, respectively, compared to the best designs used for algorithm training. Thus, this study highlights the power of combining mechanistic and machine learning models to effectively direct metabolic engineering efforts.

The C 3 -C 4 metabolite interconversion at the anaplerotic node in many microorganisms involves a complex set of reactions. C 3 carboxylation to oxaloacetate can originate from phosphoenolpyruvate and pyruvate, and at the same time multiple C 4 -decarboxylating enzymes may be present. The functions of such parallel reactions are not yet fully understood. Using a 13 C NMR-based strategy, we here quantify the individual fluxes at the anaplerotic node of Corynebacterium glutamicum, which is an example of a bacterium possessing multiple carboxylation and decarboxylation reactions. C. glutamicum was grown with a 13 C-labeled glucose isotopomer mixture as the main carbon source and 13 C-labeled lactate as a cosubstrate. 58 isotopomers as well as 15 positional labels of biomass compounds were quantified. Applying a generally applicable mathematical model to include metabolite mass and carbon labeling balances, it is shown that pyruvate carboxylase contributed 91 ؎ 7% to C 3 carboxylation. The total in vivo carboxylation rate of 1.28 ؎ 0.14 mmol/g dry weight/h exceeds the demand of carboxylated metabolites for biosyntheses 3-fold. Excess oxaloacetate was recycled to phosphoenolpyruvate by phosphoenolpyruvate carboxykinase. This shows that the reactions at the anaplerotic node might serve additional purposes other than only providing C 4 metabolites for biosynthesis.The interconversions between the C 3 metabolism and the C 4 metabolites of the tricarboxylic acid cycle function either as a replenishment of tricarboxylic acid cycle intermediates (anaplerosis) or as the initial steps of gluconeogenesis. These carboxylation and decarboxylation reactions are catalyzed by a number of enzymes (1, 2). Synthesis of oxaloacetate via carboxylation of C 3 metabolites may be catalyzed by phosphoenolpyruvate (PEP) 1 carboxylase, PEP carboxytransphosphorylase, or pyruvate carboxylase. The reverse reaction, decarboxylation of oxaloacetate, may analogously lead to PEP or pyruvate, catalyzed by PEP carboxykinase or oxaloacetate decarboxylase, respectively. NAD ϩ -or NADP ϩ -dependent malic enzyme catalyzes the reaction from malate to pyruvate. In some organisms, this enzyme is also thought to act in a pyruvate-carboxylating sense (3).To date, a full understanding of these enzymatic reactions and their functions has been hindered by a lack of knowledge about their activities in vivo. The occurrence of parallel reactions and the involvement of a set of metabolites in activity control of the enzymes prevents reliable estimations on the actual enzyme use. Moreover, it is not possible to derive quantitative data on in vivo flux rates by enzyme characterizations alone. Instead, carbon-13 labeling techniques, which employed NMR spectroscopy (for an overview, see e.g. Refs. 4 and 5) or mass spectrometry, as well as carbon-14 radiolabeling methods have been used to quantify in vivo intracellular fluxes in central metabolism including conversions between PEP, pyruvate, and oxaloacetate/malate. However, although these studies quantified the total C ...

Only 25% of bacterial membrane transporters have functional annotation owing to the difficulty of experimental study and of accurate prediction of their function. Here we report a sequence-independent method for high-throughput mining of novel transporters. The method is based on ligand-responsive biosensor systems that enable selective growth of cells only if they encode a ligand-specific importer. We developed such a synthetic selection system for thiamine pyrophosphate and mined soil and gut metagenomes for thiamine-uptake functions. We identified several members of a novel class of thiamine transporters, PnuT, which is widely distributed across multiple bacterial phyla. We demonstrate that with modular replacement of the biosensor, we could expand our method to xanthine and identify xanthine permeases from gut and soil metagenomes. Our results demonstrate how synthetic-biology approaches can effectively be deployed to functionally mine metagenomes and elucidate sequence-function relationships of small-molecule transport systems in bacteria.