To mitigate global climate change, the biorefinery platform has been implemented to largely produce fuels and chemicals for replacement of counterparts derived from fossil fuels (Saini et al., 2017; Takkellapati et al., 2018; Francois et al., 2020; Sun et al., 2021). The platform technology that manages the sustainable feedstock such as lignocellulosic biomass has many potential applications in industry. The exploitation of microbes for production of lignocellulosic ethanol (LE) presents to be a good example (Liu et al., 2019). However, the development progress of LE is afflicted by its higher production cost as compared to the petroleum price (Tuck et al., 2012). A cascade process has been suggested to produce value-added products prior to biofuels, which renders the biomass-based biorefinery economically viable (Escamilla-Alvarado et al., 2017). Enzymes, functional proteins, organic chemicals, and antibiotics are bio-based products of industrial importance. Among them, enzymes are extensively applied in the sector of biofuels, chemicals, cosmetics, detergent, food, pharmaceuticals, pulp, and textiles (Li et al., 2012; Jegannathan and Nielsen, 2013). Their global market value now reaches billions of dollars while the market volume continuously increases with industrialization.

Recognized as a biotechnology-friendly microbe, Escherichia coli has been commonly applied for the mass production of recombinant proteins. Glucose serves as the main carbon source in the culture medium. The dissimilation of glucose proceeding through glycolysis is highly efficient in E. coli. However, the presence of excessive glucose results in acetate overflow which handicaps cellular activities (Millard et al., 2021). This issue has been addressed by the strategy of lowering glucose uptake in cells, involving augmentation of the anaplerotic reaction, recruitment of inferior permeases for glucose transport, or inactivation of certain global regulators (March et al., 2002; Perrenoud and Sauer, 2005; Anda et al., 2006). An alternative method aims to improve activity of the tricarboxylic acid (TCA) cycle by decoupling the arcA/arcB-mediated control circuit (Vemuri et al., 2006). Results of these engineering approaches are generally encouraging. Nevertheless, it remains challenging to overproduce recombinant proteins in cells. The biosynthesis capacity of cells is usually overloaded by the forced expression of a recombinant protein, which leads to the adverse response known as “metabolic burden” (Oh and Liao, 2000). The shortage of cellular resources further triggers the global stress response that disables cell activities (Gill et al., 2000).

The problem of metabolic burden caused by protein overproduction can be partially solved in E. coli cultured with rich media (Oh and Liao, 2000). Yeast extract which consists of amino acids, vitamins, and trace growth elements is commonly supplemented as a major component in rich media (Bekatorou et al., 2006). However, its composition varies from batch to batch. This, in turn, renders the culturing condition and productivity of E. coli difficult to control (Potvin et al., 1997). Notice that a prerequisite for the LE production calls on the enzymatic hydrolysis of lignocellulose into fermentable sugars. Indeed, there is 20–30% of the total production cost for LE coming from hydrolytic enzymes (Singh et al., 2019). As illustrated with a hydrolytic enzyme (β-glucosidase), the techno-economic analysis reveals that glucose mainly contributes to the cost of raw materials that account for 25% of the production cost for the enzyme (Ferreira et al., 2018). Taken together, it is necessary to make more efforts toward achieving the efficient and cost-effective production of enzymes in the biorefinery sector.

In biorefinery, the processing of feedstocks which comprise starch, hemicelluloses, and oil generally leads to the associated production of proteins in the waste stream (Li et al., 2018). The estimated production of protein wastes can reach 100 million tons per year if biomass-derived fuels provide 10% of transportation fuels for the global need (Tuck et al., 2012). Therefore, it is appealing to valorize the current biorefinery process into which the protein-based production platform is integrated (Tuck et al., 2012; Li et al., 2018). Protein waste coming from various sources contains glutamate (Glu) in abundance (Lammens et al., 2012). Taking sugar beet vinasse as an example, the mass fraction of Glu reaches 63% of the total amino acids in the protein content (Deike and Mol, 1996). This apparently makes Glu promising for industrial applications owing to its renewable and sustainable nature. In this context, our previous study illustrated that Glu substituted for yeast extract in the minimal medium was useful for the enzyme production in E. coli (Chiang et al., 2020b). In contrast to complex media, minimal media which contain defined nutrient compositions are favorable for scale-up because of their easy preparation and consistent quality. In addition, evolved E. coli gains an excellent capability of metabolizing Glu (Chiang et al., 2021b). This study was therefore initiated to explore Glu as a potential substrate applied for the enzyme production. As a result, the recombinant D-hydantoinase (HDT) was highly expressed in the engineered E. coli grown on the minimal medium containing Glu that served as a carbon and nitrogen source.

Recombinant proteins have been commonly produced in E. coli using glucose. However, the adverse event of acetate overflow occurs in response to surplus glucose. Glycerol is utilized less efficiently than glucose in E. coli. This, in turn, lowers the production of acetate as a result of the induced “carbon stress-based acetate recycling pathway” (Martínez-Gómez et al., 2012). The use of glycerol is particularly useful for overproduction of aggregate-prone proteins which link to the stress response (Strandberg and Enfors, 1991). Nevertheless, an increase in glycerol utilization leads to PYR overflow in E. coli (Chiang et al., 2020a). Acetate is inexpensive but metabolized very slowly by E. coli. The acetate-based production of the single chain derivative of monellin was proven promising (Leone et al., 2015). However, the nutrient supplement is necessary to sustain growth of the strain dependent on acetate.

In this study, Glu was explored for production of the recombinant protein. BL21(DE3) strain is recognized as the most commonly used producer in the field. The recruitment of gltS* rendered the strain capable of utilizing Glu, which likely eliminates the kinetic constraint of Glu assimilation enzymes. The efficiency of the protein production was first evaluated in the strain using various substrates. Figure 2 showed that the maximum production of HDT was obtained for the strain grown on Glu. The strain metabolized various carbon sources at a different rates and displayed distinct growth profiles. Accordingly, the production rate (U/h) was calculated by dividing the HDT yield (U) by the time (h) in which the protein was sampled for the measurement. This gave the HDT productivity based on Glu, glycerol, glucose, and acetate reaching 5.5, 4.6, 3.9, and 2.3 U/h, respectively. Overall, Glu apparently serves as an excellent substrate for the protein production in E. coli.

The Glu-based production system was refined by development of BAD-HDT strain that carried the Ara-inducible T7 expression system and a genomic copy of P_T7::HDT. In Glu metabolism, the assimilation pathway leads to the production of α-KG (Figure 1). The continued oxidation of α-KG in the TCA cycle produces OAA. PEP carboxykinase (encoded by pckA) decarboxylates OAA to form PEP, which mainly serves for gluconeogenesis. It is apparent that Glu metabolism closely relies on the TCA cycle. Notice that BAD-HDT strain belongs to E. coli B strain which displays a higher activity of the TCA cycle than K12 strain (Waegeman et al., 2011). To improve the protein production in the strain, the gluconeogenic flux was directed into the oxidative route in the PP pathway by enhanced expression of zwf and pgl (Figure 1). It resulted in a 42% increase in the HDT production for BAD-HDT-1/pTH-gltS* strain (Figure 4; Table 1). PYR kinase (encoded by pykF) is subjected to the repression of Cra (encoding fructose repressor) and controls the flow direction of glycolytic flux that bifurcates at the PEP node (Bledig et al., 1996). Cra is inactivated upon binding to intermediate metabolites (i.e., glycolytic intermediates) derived from glucose (Bley et al., 2018). Our previous study illustrated that the Glu-positive strain deprived of pfkA [encoding 6-phosphofructokinase (PFK)] failed to grow on Glu (Chiang et al., 2021b). The function of PFK is responsible for phosphorylation of fructose 6-phosphate to form fructose-1,6-biphosphate. The essential role of PFK in Glu metabolism likely elevates the level of glycolytic intermediates. This, in turn, disables Cra and induces the catabolite activation of pykF. Accordingly, the further oxidization of PEP mediated by PYR kinase may restrict the gluconeogenic flux (Figure 1). This flux constraint appears to be eased by activation of the oxidative PP pathway after recruitment of zwf and pgl.

The reaction catalyzed by Asp-ammonia lyase releases ammonia in Glu metabolism (Figure 1). It was illustrated that BAD-HDT-1/pTH-gltS* strain displayed no growth defect in the Glu-containing medium with a 1/3 dosage of NH₄Cl. The result suggests that the liberated ammonium supplements the nitrogen source which is sufficient to sustain cell activities. In contrast, the administration of a 1/3 dosage of NH₄Cl caused a decrease (>50%) in the biomass of the strain grown on glucose, glycerol, or acetate. The HDT production in BAD-HDT-1/pTH-gltS* strain was further improved in the presence of high Glu (ca. 20 g/l) and low NH₄Cl (ca. 0.33 g/l) (Figure 5). However, the approach resulted in the accumulation of acetate with a concentration lower than the inhibitory level. The underlying mechanism for acetate overflow remains unclear. With an enhanced level of ppc, BAD-HDT-1/pTH-gltS*/pPC201 strain enabled reduction of acetate and production of 1.63-fold more HDT than its parent strain (BAD-HDT/pTH-gltS*). It is likely that the activity of PEP carboxylase reduces the PEP pool and provides more OAA in favor of the reaction catalyzed by citrate synthase (encoded by gltA) (Figure 1). Notice that the availability of Glu leads to formation of glutamine (Gln) through Gln synthetase (encoded by glnA). The reductive synthesis of more than 10 amino acids requires ammonia supplied by Glu and Gln (White, 2007). This apparently adds one more advantage to Glu applied for the protein production.

In summary, this preliminary study illustrated the potential of Glu for overexpressing recombinant proteins in E. coli. Glu is the first commercially-available amino acid as produced by the fermentation process, and its market size reaches 1.5 million tons per year and increasingly grows (Kumar et al., 2014). In addition, the abundance of Glu exists in the protein waste associated with the conventional biomass-treatment process. It is expected that waste streams of these production processes would provide a sustainable and renewable source of Glu. The focus of our continued work is development of a streamlined process to recover Glu from the protein waste. Accordingly, the developed approach based on Glu may open up a new avenue for the enzyme biorefinery platform.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

C-JC and Y-PC conceived the idea and wrote the manuscript. M-CH and TT performed the experiment. All authors contributed to the article and approved the submitted version.

This work was co-supported by China Medical University (CMU111-MF-79) and the Ministry of Science and Technology (MOST 108-2221-E-035-052-MY3), Taiwan.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.