Discussion
In the last decades of the 20th century, E.coliBL21 (DE3) has become the preferred host for recombinant protein production. However,some recombinant proteins may impose a high metabolic burden or lead to toxicity in the host cell, which may result in reduced growth rate, low final cell density, and even cell death (Bhattacharya & Dubey, 1995). After 24h, GDH expression inhibited cell growth and induced severe autolysis (Fig. 1A and B). For the expression of toxic membrane protein, a common strategy is to decrease the expression of the toxin protein by governing the expression of T7 RNAP, such as C41 (DE3) and Lemo21 (DE3). Notably, the C41 (DE3) strain weaken the lacUV5 promoter by recA-dependent recombination with the lac promoter (Susan et al., 2015), indicating that the strength of the lac promoter may not be optimal. However, the C41 (DE3) strain could not effectively improve the expression of GDH compared to BL21 (DE3), which implied that the toxic effect of GDH is different from that of toxic membrane proteins. Most importantly, the autolysis phenotype appeared during the expression of a variety of proteins, which necessitates shortening the fermentation time, and eventually reduces the protein yield. The host mutation of BL21 (DE3-lac1G) changes one base in the lac-1A promoter, which had an enormous positive impact on autolysis, ie., a high rate of 89.55 U/mL/h was obtained at 43h in comparison of that of BL21 (DE3) with 3 U/mL/h. Taken together, our results proved that the new strain BL21 (DE3-lac1G) strain can effectively suppressing PCD and maintain plasmid stability, which will make it a popular host for protein production, especially for proteins requiring a longer fermentation time or maturation with after processing.
The production of foreign proteins often imposes a metabolic burden on the host by triggering local and global cellular stress responses (Bhattacharya & Dubey, 1995). In particular, this metabolic burden often manifests as an increase of the energy demand or maintenance energy requirement such as amino acids, ribosomes or other precursors (Mairhofer et al., 2013). The metabolic burden on host cells can be improved by making right choices for promoter (Pasini et al., 2016), plasmid copy number (Flores et al., 2004), and removal of codon bias (Rahmen et al., 2015). Moreover, optimizing environment conditions also can reduced metabolic burden, such as addition of amino acids or using complex medium (Fong & Wood, 2010). As a first measure, if the growth rate of recombinant strain is inhibited, then two causes may explain the phenotype: gene toxicity and basal expression of the toxic mRNA/protein. Protein toxicity refers to the toxicity of the target protein itself, such as its enzyme activity or cell damage due to misfolded aggregates (Binepal et al., 2012). On the other hand, studies have reported that excessive T7 RNAP itself can be lethal to cells (Davanloo et al., 1984). Moreover, it was found that the toxicity caused by mRNA may be neither related to plasmid abundance nor to the abundance of the encoded mRNA (Mittal et al., 2018). Previous studies also proved that the toxicity of certain gene was only dependent on transcription but independent of protein translation (Li & Rinas, 2020). In this study, a GDH mutant with dramatically reduced activity and a construct without RBS proved that autolysis was not caused by the activity or amount of protein, so it is likely that the autolysis is induced by certain mRNA elements. In this study, the T7 RNAP expression from lac-1G promoter was found to be higher than from lac-1A promoter, but lower than from lacUV5. Consequently, the rate of protein overexpression in BL21 (DE3-lac1G) was manipulated at a will level, not only transcription with a higher level but also no more than the host tolerance during the whole fermentation stage. A recent study of a plasmid-driven T7 (PDT7) system also suggested that high expression of T7 RNAP affected cell metabolism and led to toxicity and instability (Tan & Ng, 2020). In fact, the T7 RNAP itself could be not toxic, but when it combined with a strong promoter, it can induce severe growth defects. A excessively strong T7 RNAP system robbed energy from the basal metabolism, which is the major reason of T7 RNAP toxicity (Tan & Ng, 2020). By inhibiting endogenous RNAP or reducing parts of non-essential proteome production, it is possible to balance cell growth and recombinant protein production on resources allocation (Kim et al., 2019). Recently, the evolved T7 phage RNAP inhibitor Gp2 was used in BL21 (DE3) to decouple recombinant protein production from cell growth, which enhanced protein yields up to 3.4-fold (Stargardt et al., 2020). Moreover, the resources can be selectively allocated for transcription or translation of target genes by orthogonal molecular elements (Darlington et al., 2018; Segall-Shapiro et al., 2014), which were beneficial to reduce the metabolic burden of the host cells. This study proved that the key to regulating autolysis and protein overexpression is controlling the expression of T7 RNAP, which plays the role of the main on-off “switch” in the pET system. Moreover, controlling the rate of transcription of T7 RNAP can mitigate the metabolic burden effectively and easily. The new host strain of BL21 (DE3-lac1G) can effectively produce recombinant protein without affecting growth.
Under excessive stress, the repair mechanisms will be overwhelmed and cells will undergo programmed cell death (PCD). This program offers no direct advantage to individual cells, but could benefit its siblings by releasing nutrients for other cells in the colony or preventing the spread of viruses (Tanouchi et al., 2013). The death involved PCD pathway is meditated by an intracellular program (Nagamalleswari et al., 2017), mainly including the toxin-antitoxin system, holin-antiholin system and ALD pathway. The membrane integrity of BL21 (DE3) was impaired when overexpressing GDH. More importantly, the expression level of PCD markers including recA, mazEF, and yohJK were showed to be up-regulated at 43h compared to 24h (Fig. 2A), which suggested that BL21(DE3) suffered PCD at 43h. In general, deletion of the key genes involved in the PCD pathway can restore cell viability. For example, periplanetasin-2 was proven to induce apoptosis-like death inE.coli according to physiological changes, which was proven when the antibacterial activity of periplanetasin-2 was decreased by deletion of recA (Lee et al., 2019a). However, deletion of three key genes did not restore viability in the GDH production process, indicating that cell death of BL21 (DE3) does not proceed only through these three general bacterial PCD pathway. In the new strain BL21 (DE3-lac1G), the lower protein expression rate did not exceed the cellular metabolic capacity, and therefore could not induce PCD. Nevertheless, the common characteristics of proteins that cause PCD and how it is promoted needs further study.
In the future, this novel autolysis system can be used in secretory protein production. The production of extracellular proteins has distinct advantages, such as simplifying the disruption of the cell wall and purification processes (Su et al., 2013). Previous studies applied holins and endolysins to promote cell lysis, combined with various inducible promoters to prevent cell lysis before sufficient cell growth (Choi & Lee, 2004). However, expression of lytic proteins still uses valuable cellular energy resources. Inspired by this study, the promoter of T7 RNAP can be designed to let cells overexpress too much of the target protein at a desired time-point induce autolysis, so that the host produces the target protein from beginning to end, without relying on exogenous lysing proteins. In addition, four promoters with different strength can form a series of protein expression hosts with different T7 RNAP expression levels, providing a variety of host choices. Moreover, the promoter strength of lacUV5-1A is 2.68 times higher than that of the strong promoter lacUV5, which indicated that lacUV5-1A promoter can satisfy the requirements of high expression of target genes in metabolic engineering in the future.