Pareto optimal metabolic engineering for the growth‐coupled overproduction of sustainable chemicals

Abstract

Our research aims to help industrial biotechnology develop a sustainable economy using green technology based on microorganisms and synthetic biology through two case studies that improve metabolic capacity in yeast models Yarrowia lipolytica (Y. lipolytica) and Saccharomyces cerevisiae (S. cerevisiae). We aim to increase the production capacity of beta-carotene (β-carotene) and succinic acid, which are among the highest market demands due to their versatile use in numerous consumer products. We performed simulations to identify in silico ranking of strains based on multiple objectives: the growth rate of yeast microorganisms, the number of used chromosomes, and the production capability of β-carotene (for Y. lipolytica) and succinate (for S. cerevisiae). Our multiobjective optimization methodology identified notable gene deletions by searching a vast solution space to highlight near-optimal strains on Pareto Fronts, balancing the above-cited three objectives. Moreover, preserving the metabolic constraints and the essential genes, this study produced robust results: seven significant strains of Y. lipolytica and seven strains of S. cerevisiae. We examined gene knockout to study the function of genes and pathways. In fact, by studying the frequently silenced genes, we found that when the GPH1 gene is knocked out in S. cerevisiae, the isocitrate lyase enzyme is activated, which converts the isocitrate into succinate. Our goals are to simplify and facilitate the in vitro processes. Hence, we present strains with the least possible number of knockout genes and solutions in which the genes are turned off on the same chromosome. Therefore, we present results where the constraints mentioned above are met, like the strains where only two genes are switched off and other strains where half of the knockout genes are on the same chromosome. This study offers solutions for developing an efficient in vitro mutagenesis for microorganisms and demonstrates the efficiency of multiobjective optimization in automatizing metabolic engineering processes.

The Pareto principle: To what extent does it apply to resource acquisition in stable microbial communities and thereby steer their geno−/ecotype compositions and interactions between their members?

“…Distinct from but related to the Pareto principle is the concept of Pareto Optimality/Efficiency, which describes the situation in which no further improvement in a particular process can be achieved by change; that is, improvement of one parameter must lead to a worsening of another. Also used initially in the field of economics, Pareto Optimality has since been applied to a range of other fields, including microbiology (Amaradio et al, 2022; Fuentes et al, 2021; Schuech et al, 2019; Sheftel et al, 2013; Shoval et al, 2012).…”

“…Since the Pelz et al (1999) publication, there have been a number of studies that indicate that the Pareto principle may apply widely to stable microbial communities, often in the context of metabolic trade‐offs (e.g., De Vrieze & Verstraete, 2016; Fuentes et al, 2021; Marzorati et al, 2008; Mertens et al, 2005; Schink et al, 2022), but also in terms of the relationship between cell morphology and function (Schuech et al, 2019; Sheftel et al, 2013; Shoval et al, 2012), the composition of human microbiome communities (Heinken & Thiele, 2015a, 2015b), and the biotechnological use of microbes for the production of chemicals (Amaradio et al, 2022).…”

“…In general, nature does not favour the absolute dominance of one genotype over all others and it is common in biology that indirect/external forces act to regulate the frequency/ratio of organisms at a given site (Cray et al, 2013), witness for example the diversity and spacing of trees in forests resulting from conspecific negative density dependence (Johnson et al, 2012). Marzorati et al, 2008;Mertens et al, 2005;Schink et al, 2022), but also in terms of the relationship between cell morphology and function (Schuech et al, 2019;Sheftel et al, 2013;Shoval et al, 2012), the composition of human microbiome communities (Heinken & Thiele, 2015a, 2015b, and the biotechnological use of microbes for the production of chemicals (Amaradio et al, 2022). And there are other examples of stable microbial communities that might be Pareto conform, such as the kefir grains in the fermented milk product kefir (https:// sitn.hms.harvard.edu/flash/2022/cooperation-vs-compe tition-microbiome-diversity-and-interactions/; Marsh et al, 2013;Prado et al, 2015) that are dominated by Lactobacillus kefiranofaciens; other fermented products; the human vaginal microbiota that is in many cases dominated by lactobacilli (Chen et al, 2021;Ravel et al, 2011); the human skin microbiota that is dominated by Cutibacterium acnes (Grice & Segre, 2011;Rozas et al, 2021); and geothermal vent communities.…”

Comparative Transcriptome Analysis Revealing the Enhanced Volatiles of Cofermentation of Yeast and Lactic Acid Bacteria on Whole Wheat Steamed Bread Dough

“…In contrast, GLT1-encoded glutamate synthase could facilitate the glutamate synthesis. 35 The GDH gene was significantly downregulated, while the GLT1 gene was significantly upregulated in this study, resulting in the accumulation of glutamate. In addition, PDC1 related to αketo acid conversion was upregulated by 2.6-fold, favoring the subsequent conversion of α-keto acid to aldehydes.…”

L‐lactate production in engineered Saccharomyces cerevisiae using a multistage multiobjective automated design framework

“…Such models can thus open the possibility for more effective and cheap design flows in the future. In fact, computational tools are now commonly used to enhance the production of desired compounds via the discovery of efficient metabolic routes, or by unveiling novel pathways able to produce chemicals that could previously not been obtained from the specific organism (Amaradio et al, 2022). Different modification techniques are available for the purpose, including gene deletions or regulations; but the modeler can also focus its attention on other variables that affect the cell productivity, such as its environment (Long et al, 2015).…”