Comparative genomics and experimental evolution of Escherichia coli BL21(DE3) strains reveal the landscape of toxicity escape from membrane protein overproduction

Abstract

Achieving sufficient yields of proteins in their functional form represents the first bottleneck in contemporary bioscience and biotechnology. To accomplish successful overexpression of membrane proteins in a workhorse organism such as E. coli, defined and rational optimization strategies based on an understanding of the genetic background of the toxicity-escape mechanism are desirable. To this end, we sequenced the genomes of E. coli C41(DE3) and its derivative C43(DE3), which were developed for membrane protein production. Comparative analysis of their genomes with those of their ancestral strain E. coli BL21(DE3) revealed various genetic changes in both strains. A series of E. coli variants that are able to tolerate transformation with or overexpression of membrane proteins were generated by in vitro evolution. Targeted sequencing of the evolved strains revealed the mutational hotspots among the acquired genetic changes. By these combinatorial approaches, we found non-synonymous changes in the lac repressor gene of the lac operon as well as nucleotide substitutions in the lacUV5 promoter of the DE3 region, by which the toxic effect to the host caused by overexpression of membrane proteins could be relieved. A mutation in lacI was demonstrated to be crucial for conferring tolerance to membrane protein overexpression.

Figure 1. Genomic variations in E. coli C41(DE3) and C43(DE3) as compared to the BL21(DE3) backbone sequence in a circular representation.

Mutations found in both C41(DE3) and C43(DE3) are depicted in red, C41(DE3) in green, and C43(DE3) in blue, respectively. Small-scale changes like SNPs or small DIPs are indicated with solid lines, and IS element-mediated large scale insertions or deletions with triangles.

Figure 2. Differences in the promoter sequences that control the transcription of the T7 RNA polymerase gene between E. coli BL21(DE3) and fifteen mutant derivatives that overcame the transformation toxicity induced by a membrane protein-encoding gene.

Three strains in the mutant group BL1 had the same promoter sequence as C41(DE3) and C43(DE3). Nine BL2 mutants had reverted the lacUV5 promoter sequence to that of the wild-type lac promoter. Three BL3 strains showed no difference from the lacUV5 promoter of BL21(DE3).

Figure 3. Expression capacity of E. coli C41(DE3), C43(DE3), C41(DE3)lacIV192F, and C43(DE3)lacIF192V to produce the E. coli F-ATPase subunit b.

(a) Phenotypic comparison of Ecb expression in C41(DE3), C43(DE3), and C41(DE3)lacIV192F in the absence (-IPTG) or presence (+IPTG) of IPTG. C41(DE3), C43(DE3), and C41(DE3)lacIV192F contained pMW7(Ecb-GFP). (b) Relative expression levels of the atpF gene in C41(DE3), C43(DE3), C41(DE3)lacIV192F, and C43(DE3)lacIF192V containing pMW7(Ecb-GFP). After IPTG induction, the cultures were collected for RNA extraction. Quantification of the mRNA expression level was performed using qRT-PCR and the comparative critical threshold (2^−ΔΔCT) method (Livak and Schmittgen, 2001). The 16S rRNA gene was used as an internal control. The transcripts at the four time points (shown on top) relative to the transcript levels in C41(DE3) at 0 hr were quantified. Error bars indicate the standard deviations of triplicate reactions. (c) Flow cytometric analysis of GFP fluorescence in C41(DE3) (black), C43(DE3) (blue), C41(DE3)lacIV192F (red) and C43(DE3)lacIF192V (orange) overexpressing Ecb-GFP. After 3 hours of IPTG induction, the cells were collected for detection of GFP expression. C41(DE3) without the plasmid was used as the negative control.

Figure 4. Structural analysis of the V192F mutation in the LacI protein.

The model shows the steric hindrance caused by V192F, which results in changes to the sugar- or IPTG-binding pocket. Steric clashes are depicted as red disks. The model was illustrated with PyMOL.

Figure 5. Overview of the in vitro evolution of E. coli strains overexpressing membrane proteins.

A scheme to generate BL and CL strains that are capable of overcoming the toxicity from transformation with or overexpression of Ecb, respectively, are shown in the upper part of the figure. The selection of C41(DE3) from BL21(DE3), and C43(DE3) from C41(DE3), was conducted previously. The mutational hotspots in the evolved strains and the probable mechanisms producing them are illustrated at the bottom. Mutational hotspots were located in the lacUV5 promoter of the DE3 region and in the lacI gene of the lac operon, which were identified as decisive mutations that made E. coli cells tolerant to the overexpression of membrane proteins (Wagner et al. and this study). The mutations in the lacUV5 promoter in the DE3 region result from homologous recombination with the similar lac promoter in the lac operon.