1. Introduction
Yeast has been widely used as a “cell factory” in industrial fermentation processes to produce a wide range of valuable products, including organic acids that are used extensively in manufacturing, pharmaceutical, cosmetic, food, textile and chemical industries (Álvarez-Chávez et al., 2012; Chen et al., 2013; de Jong et al., 2014; Gonzalez-Garcia et al., 2017; Hong and Nielsen, 2012). Compared to conventional chemical methods for the production of organic acids based on fossil fuel reserves, microbial production is an attractive approach due to several advantages including sustainability, less environmental pollution and cost-effectiveness (Sauer et al., 2008; Steen et al., 2010). Unlike other hosts that are recalcitrant to genetic manipulation (Kiatpapan and Murooka, 2002; Zhuge et al., 2013), baker’s yeast (Saccharomyces cerevisiae ) is an ideal organism to discover new gene targets for productivity enhancement, because it is a model eukaryotic organism with high-resolution genomic data. Moreover, the tolerance of yeast to low pH enables the production of organic acids in their protonated forms, reducing the costs of downstream recovery and purification after fermentation. Due to the economic, environmental and medical importance of organic acid production by yeast, advanced metabolic engineering and synthetic biology technologies have been applied to engineer yeast for improved production of different high-value organic acids, such as lactic acid (Ishida et al., 2005), succinic acid (Otero et al., 2013), para-hydroxybenzoic acid (Williams et al., 2015), 3-hydroxypropionic acid (Borodina et al., 2015) and muconic acid (Curran et al., 2013).
Yeast is also an attractive host for the production of propionic acid (PA) that is commonly used as a food preservative and a chemical intermediate, since PA can be formed as a by-product of yeast fermentation (Eglinton et al., 2002). However, PA is toxic to yeast, especially at relatively low concentrations, causing an important problem of tolerance engineering in yeast PA production. Fortunately, Xu et al. (2019) demonstrated significant improvements in yeast tolerance to PA using adaptive laboratory evolution (ALE), a powerful tool in the field of metabolic engineering for the development of superior industrial microbial strains (Almario et al., 2013; Gonzalez-Ramos et al., 2016; Kildegaard et al., 2014).
ALE experiments are lab-intensive and time-consuming however, requiring evaluation of growth kinetics of intermediate populations and numerous candidate strains to select an ideal strain with improved phenotypes. The evolution process might be performed over hundreds of generations, and the traditional growth test based on optical density (OD) measurements must be conducted over three repetitions for each strain or population for each population or strain. Moreover, the process lacks the ability to track the growth of yeast at a single-cell level, and cannot consider cell size, morphology and viability that may change during growth. Thus, cell-to-cell variations are obscured and the ability to screen and select single cells with desired characteristics (e.g., high growth rate, high tolerance to acids and high secretion of valuable bio-products) is limited.
In order to address these limitations, an alternative approach is required to quantitatively track the growth of individual cells within a population without perturbation and allows parallel, high-throughput assessment at a single-cell level. Microfluidics can compartmentalize single cells within monodisperse picolitre-sized droplets in a cost-effective and high-throughput process, for example, screening of 5 × 107 individual reactions requires only 150 µL of reagents and seven hours at an estimated cost of only a few dollars, as demonstrated by Agresti et al. (2010). Over the past decades, droplet microfluidics has enabled single-cell analysis for a wide range of applications across biological science, biomedicine and biochemistry (Agresti et al., 2010; Brouzes et al., 2009; Yu et al., 2018). This is because (1) the extracellular environments are accurately mimicked (Hosokawa et al., 2017; Liu et al., 2020); (2) the genotype-phenotype linkages are established at a single-cell level (Bowman and Alper, 2020; Fischlechner et al., 2014; Li et al., 2018a, 2018b); (3) the miniaturized confinement improves the detection limit (Agresti et al., 2010; Zhu et al., 2012); and (4) massive parallel analysis can be conducted to probe cellular heterogeneity (Headen et al., 2018; Hindson et al., 2011; Klein et al., 2015; Ostafe et al., 2014; Zinchenko et al., 2014).
In this study, we quantitatively tracked the growth of single yeast cells under varying conditions by using monodisperse microdroplets. In order to demonstrate the versatility of the microdroplet platform, we used two species of yeast, Saccharomyces cerevisiae (S. cerevisiae ) and Pichia pastoris (P. pastoris ), and a total of four strains, wild-type S. cerevisiae strain (CEN.PK113-7D), the PA evolved mutant S. cerevisiae strain (PA-3), GFP-tagged S. cerevisiae strain (CEN.PK2-1C-GFP) and GFP-tagged P. pastoris strain (CBS7435-GFP). The effects of organic acids, PA and AA, at different concentrations on the growth of yeast at the single-cell level were studied, as well as the effect of K-ions on PA tolerance in yeast. The calculated specific growth rate (μ) of single yeast grown in microdroplets was effectively identical to that for cells in bulk cultures at a pH of 3.5, and yeast cells maintained high viability in microdroplets after 48 hours of culture.