In order to improve the production of various AMPs, it needs to construct the library of BL21 (DE3)-derived variant strains. For target AMP, the most suitable expression host could be screened from the library in only three days with high throughput fluorescence screening, thus realizing "Customized production of specific AMPs". Based on this design, our project contain two aspects: construction of library and application of library.
1. T7 RNAP was selected as the core of the modified pET expression system
The pET expression system is the most representative AMPs expression systems. It is reported that the transcriptional capacity of T7 RNAP is approximately 8-fold greater than that of the endogenous RNAP of E. coli [1-3]. However, BL21 (DE3) cannot efficiently produce AMPs due to the natrual toxic effect of AMPs. Constructing a more appropriate expression host is a good choice for high–level expression of certain hard-to-express proteins. A series of BL21 (DE3)-derived variant strains, including C41(DE3), C43(DE3), BL21 (DE3-lac1G), Lemon21(DE3), and Mutant56(DE3), were developed for toxic protein production. It is worth mentioning that the rationale behind these variant strains is that the expression or activity of the T7 RNAP was reduced, which in turn produced less T7 RNAP corresponding to the target recombinant protein. For AMP synthesis, this will reduce the host toxicity caused by rapid synthesis of AMPs, and facilitate AMPs production. Therefore, we decide to modify the expression of T7 RNAP as the key of constructing better expression host (Fig. 1).
2. The RBS sequences of T7RNAP was modified to creat the library of BL21 (DE3)-derived variant strains
The intensity of T7 RNAP expression can be regulated at different levels to solve the above issues. (I) At the level of transcription, the lacUV5 promoter of T7RNAP was weakened in many variant strains, including the lacWT promoter of C41 (DE3)/C43(DE3) and the lac-1G promoter of BL21 (DE3-lac1G). (II) At the level of translation, it can be carried out by changing the ribosomal binding site (RBS) sequences or resolving the codon bias of target gene. Among the above-mentioned ways, the transcriptional and translational regulations are commonly used. The promoter involves many regulatory elements, such as -10 region and -35 region, making it more difficult to adjust the intensity of T7 RNAP. In contrast, by altering the RBS sequence, it takes only a few number of base mutations to regulate the translation strength (Fig. 2). For instance, nucleotide changes to the RBS can affect protein production levels by as much as 600-fold. Therefore, in our project, creating the BL21 (DE3)-derived variant strains library with different RBS sequences of T7RNAP, can quickly provide the most suitable host for recombinant protein production, rather than requiring the construction of multiple variants.
3. Using the Base Editor to construct the library of BL21 (DE3)-derived variant strains
CRISPR-mediated DNA base editor is a promising tool that can yield point mutations at desired sites without generating a DNA double-strand break or requiring a donor DNA template. It has been reported that the base editor can be applied to Corynebacterium glutamicum, Bacillus subtilis, Saccharomyces cerevisiae and E. coli MG1655. The cytosine base editor (CBE), consisting of dSpCas9, CDA, UGI and LVA, has been tested in E. coli MG1655, with a high editing efficiency of 87.2%. Therefore, the CBE was designed to build the RBS library of T7 RNAP in BL21 (DE3) firstly (Fig. 3). To facilitate the editing of the RBS sequence using CBE, a tailored RBS sequence of 5'-GGGGGGGG-3' was integrated into the corresponding site of the T7 RNAP, yielding a strating strain of BL21 (DE3)-RBS8G. However, the editing range was limited to 4-5 bases in the distal part of RBS8G. To extend the editing range, two different sgRNA expression frames were tandemly linked, allowing GGGGGGGG to be covered, resulting in a more diverse editing outcome.
4. Using the CRISPR/Cas9 to construct the library of BL21 (DE3)-derived variant strains
Under the guide of sgRNA, the Cas9 protein can generate the double-strand breaks, which can be repaired using donor DNA by homologous recombination in BL21 (DE3). Therefore, the new different RBS sequences can be introduced in the Donor DNA, then it can be integrated into the genome to produce a library of BL21 (DE3)-derived variant strains (Fig. 4).
5. The application of the library of BL21 (DE3)-derived variant strains
Based on the BL21 (DE3)-derived variant strains library, the high throughput host screening platform was constructed to screening the producer with the high recombinant protein titer. In this system, the target recombinant protein was fused with the eGFP and was transferred into the BL21 (DE3)-derived variant strains library, then the best host was screened by fluorescence intensity in 96-well plates. Using this platform, it takes only three days to obtain the highest-expression host (Fig. 5).
 Iost I, Guillerez J, Dreyfus M. Bacteriophage T7 RNA polymerase travels far ahead of ribosomes in vivo. Journal of bacteriology. 1992; 174: 619–622.
 Carmignotto GP, Azzoni AR. On the expression of recombinant Cas9 protein in E. coli BL21(DE3) and BL21(DE3) Rosetta strains. Journal of biotechnology. 2019; 306: 62–70.
 Rostinawati T, Paramita NG, Wicaksono IA, Sriwidodo SS, Yusuf M, Subroto T. Characterization of Anti-HER2 scFv Gene Expression as Intracellular Protein in Escherichia coli BL21 (DE3). International journal of pharmaceutical investigation. 2020; 10: 117–121.
 Walker BM. Over-production of Proteins in Escherichia coli: Mutant Hosts that Allow Synthesis of some Membrane Proteins and Globular Proteins at High Levels. Journal of molecular biology. 1996; 260: 289–298.
 Sun XM, Zhang ZX, Wang LR, Wang JG, Liang Y, Yang HF. Tao RS, Jiang Y, Yang JJ, Yang S. Downregulation of T7 RNA polymerase transcription enhances pET-based recombinant protein production in Escherichia coli BL21 (DE3) by suppressing autolysis. Biotechnology and bioengineering. 2021; 118: 153–163.
 Schlegel S, Löfblom J, Lee C, Hjelm A, Klepsch M, Strous M, Drew D, Slotboom DJ, De Gier JW. Optimizing membrane protein overexpression in the Escherichia coli strain Lemo21(DE3). Journal of molecular biology. 2012; 423: 648–659.
 Baumgarten T, Schlegel S, Wagner S, Löw M, Eriksson J, Bonde I, Herrgård MJ, Heipieper HJ, Nørholm MHH, Slotboom, DJ, De Gier JW. Isolation and characterization of the E. coli membrane protein production strain Mutant56(DE3). Scientific reports. 2017; 7: 1–14.
 Mutalik VK, Guimaraes JC, Cambray G, Lam C, Christoffersen MJ, Mai QA, Tran AB, Paull M, Keasling JD, Arkin AP, Endy D. Precise and reliable gene expression via standard transcription and translation initiation elements. Nature Methods. 2013; 10: 354–360.
 Wang Y, Cheng H, Liu Y, Wen X, Zhang K, Ma Y. In-situ generation of large numbers of genetic combinations for metabolic reprogramming via CRISPR-guided base editing. Nature communications. 2021; 12(1): 1-12.
 Gong G, Zhang Y, Wang Z, Liu L, Shi S, Siewers V, Yuan Q, Nielsen J, Zhang X, Liu Z. GTR 2.0: GRNA-tRNA Array and Cas9-NG Based Genome Disruption and Single-Nucleotide Conversion in Saccharomyces cerevisiae. ACS synthetic biology. 2021; 10: 1328–1337.
 Zhao D, Li J, Li S, Xin X, Hu M, Price MA, Rosser SJ, Bi C, Zhang X. Glycosylase base editors enable C-to-A and C-to-G base changes. Nature Biotechnology. 2021; 39: 35–40.