CRISPR interference-guided multiplex repression of endogenous competing pathway genes for redirecting metabolic flux in Escherichia coli

Abstract

Background: Multiplex control of metabolic pathway genes is essential for maximizing product titers and conversion yields of fuels, chemicals, and pharmaceuticals in metabolic engineering. To achieve this goal, artificial transcriptional regulators, such as clustered regularly interspaced short palindromic repeats (CRISPR) interference (CRISPRi), have been developed to specifically repress genes of interest.

Results: In this study, we deployed a tunable CRISPRi system for multiplex repression of competing pathway genes and, thus, directed carbon flux toward production of molecules of interest in Escherichia coli. The tunable CRISPRi system with an array of sgRNAs successfully repressed four endogenous genes (pta, frdA, ldhA, and adhE) individually and in double, triple, or quadruple combination that are involved in the formation of byproducts (acetate, succinate, lactate, and ethanol) and the consumption of NADH in E. coli. Single-target CRISPRi effectively reduced the amount of each byproduct and, interestingly, pta repression also decreased ethanol production (41%), whereas ldhA repression increased ethanol production (197%). CRISPRi-mediated multiplex repression of competing pathway genes also resulted in simultaneous reductions of acetate, succinate, lactate, and ethanol production in E. coli. Among 15 conditions repressing byproduct-formation genes, we chose the quadruple-target CRISPRi condition to produce n-butanol in E. coli as a case study. When heterologous n-butanol-pathway enzymes were introduced into E. coli simultaneously repressing the expression of the pta, frdA, ldhA, and adhE genes via CRISPRi, n-butanol yield and productivity increased up to 5.4- and 3.2-fold, respectively.

Conclusions: We demonstrated the tunable CRISPRi system to be a robust platform for multiplex modulation of endogenous gene expression that can be used to enhance biosynthetic pathway productivity, with n-butanol as the test case. CRISPRi applications potentially enable the development of microbial "smart cell" factories capable of producing other industrially valuable products.

Keywords: CRISPR interference; Endogenous gene; Escherichia coli; Multiple gene knockdown; n-Butanol.

Fig. 1

CRISPRi system design and construction for reducing byproduct formation in engineered E. coli producing n-butanol. a Schematic representation of the n-butanol-production pathway. The reconstituted n-butanol-production pathway consists of five enzymes involved in the six-step synthesis of n-butanol from acetyl-CoA. Glucose and glycerol mainly used for n-butanol production in E. coli are redirected into succinate, lactate, acetate, and ethanol production during glycolysis. b A plasmid encoding the five genes of the reconstructed n-butanol pathway. The expression of n-butanol-pathway genes was controlled by a constitutive lac promoter (lacP′). c The CRISPRi plasmid harboring an sgRNA array consisting of four sgRNAs targeting pta, frdA, ldhA, and adhE genes in E. coli. The CRISPRi system consisted of a dCas9 protein and sgRNAs governed by an l-rhamnose-inducible promoter (P_rhaBAD) and a J23119 constitutive promoter (P_J23119), respectively. P, F, L, and A are sgRNAs targeting endogenous pta, frdA, ldhA, and adhE gene, respectively. AdhE, aldehyde/alcohol dehydrogenase; AtoB, acetyl-CoA acetyltransferase; Cam^R, chloramphenicol-resistance gene; Crt, crotonase; DHAP, dihydroxyacetone phosphate; FBP, fructose 1,6-bisphosphate; FrdA, fumarate reductase; G3P, glyceraldehyde 3-phosphate; Hbd, 3-hydroxybutyryl-CoA dehydrogenase; Kan^R, kanamycin resistance gene; LdhA, lactate dehydrogenase; PEP, phosphoenolpyruvate; Pta, phosphate acetyltransferase; Ter, trans-enoyl-CoA reductase

Fig. 2

Schematic description of CRISPRi-mediated repression of endogenous genes involved in byproduct formation. a Genomic structure of four endogenous genes in E. coli. The ldhA and adhE genes are monocistronic, whereas frdA and pta reside in frdABCD and ackA-pta operon, respectively, in E. coli. In the ackA-pta operon, pta is controlled by a separate promoter in addition to the ackA promoter. b The CRISPRi plasmid harboring an sgRNA targeting the pta, frdA, ldhA, or adhE gene in E. coli. c A synthetic fluorescence-based gene-reporter plasmid containing a GFP-encoding gene. Binding sites for each dCas9-sgRNA complex were inserted between the constitutive J23100 promoter (P_J23100) and the gfp gene in the pREGFP3 plasmid. The expression of the gfp gene was controlled by the constitutive J23100 promoter (P_J23100). d, e In vivo fluorescence assay reporting CRISPRi-mediated GFP repression. Two plasmids (pSECRi and pREGFP3) were introduced into E. coli, and expression of the dCas9 protein was induced by 4 mM l-rhamnose. f mRNA levels of CRISPRi target genes in E. coli. The pSECRi plasmid was transformed into E. coli cells, and real-time PCR was performed to determine the mRNA levels of each gene (pta, frdA, ldhA, or adhE). The expression ratios of GFP were calculated as $expression(\%)=\frac{RF{U}_{xv}/O{D}_{xv}}{RF{U}_{null}/O{D}_{null}}\times 100,$ where RFU and OD are relative fluorescence units and optical density values at 600 nm, respectively. The subscript xv designates the tested cells harboring the pSECRi plasmid in the presence of l-rhamnose, whereas null indicates a control with the same pSECRi plasmid in the absence of l-rhamnose

Fig. 3

Construction of sgRNA arrays using a restriction digest-based modular assembly. Plasmids pSECRi-P, -F, -L, and -A encode individual sgRNAs that are flanked by AgeI and XmaI on the 5′ and 3′ end, respectively. These flanking restriction sites in all pSECRi plasmids allow us to connect any two sgRNAs by ligating a plasmid backbone digested with NcoI and AgeI to an insert fragment obtained by digestion with NcoI and XmaI. The resulting plasmid (pSECRi-PF, or -LA) harbors a two-sgRNA array that again is flanked by AgeI and XmaI on the 5′ and 3′ end, respectively. We should mention that AgeI and XmaI have compatible cohesive ends and ligation of both DNA fragments digested with AgeI and XmaI generates a new restriction site that is not cleavable by both restriction enzymes. A two-sgRNA array can be joined to the array by ligating a plasmid backbone digested with NcoI and AgeI to an insert digested with NcoI and XmaI, resulting in a pSECRi-PFLA. All sgRNAs are under the control of J23119 promoter (P₁–P₄)

Fig. 4

Comparison of the effect of CRISPRi-mediated multiplex gene repression on n-butanol production in E. coli. DH5α cells harboring pAB-HCTA and each pSECRi plasmid were grown in TB medium supplemented with 20 g/L glycerol and a 4 mM l-rhamnose or b various concentrations of l-rhamnose at 37 °C for 48 h. n-Butanol levels were determined by GC. The fold increase of n-butanol production was calculated as $n\text{-}butanol(fold)=\frac{BtO{H}_{xv}}{BtO{H}_{null}}\times 100,$ where BtOH is n-butanol concentration. The subscript xv designates the tested cells harboring the pAB-HCTA and pSECRi plasmid in the presence of l-rhamnose, whereas null indicates a control with the same two plasmids in the absence of l-rhamnose. Data represent the averages of three biological cultures, and error bars show the standard deviation (SD)

Fig. 5

Reductions in n-butanol-derived byproducts and screening of various E. coli strains for n-butanol production. Escherichia coli DH5α cells containing pAB-HCTA(Cam^R) or pABA-HCTA(Amp^R) were grown in TB medium supplemented with 20 g/L glycerol at 37 °C for 48 h. a, b Metabolic products, including n-butanol and its derivatives, were determined by GC. The fold increase of n-butanol production was calculated as $n\text{-}butanol(fold)=\frac{BtO{H}_{amp}}{BtO{H}_{cam}}\times 100,$ where BtOH is n-butanol concentration. The subscript amp designates the tested cells harboring the pABA-HCTA, whereas cam indicates a control with the pAB-HCTA plasmid. c To screen the best n-butanol producer among various E. coli strains, three different E. coli strains (DH5α, MG1655, and BW25113) were examined. Two plasmids, pSECRi-PFLA carrying a quadruple sgRNA array and pABA-HCTA encoding heterologous n-butanol-pathway genes, were introduced into each E. coli strain, which was grown in TB medium supplemented with 20 g/L glycerol and 4 mM l-rhamnose at 37 °C for 48 h. n-Butanol levels were determined by GC. The fold increase of n-butanol production was calculated as $n\text{-}butanol(fold)=\frac{BtO{H}_{xv}}{BtO{H}_{dh}}\times 100,$ where BtOH is n-butanol concentration. The subscript xv designates the tested cells harboring the pABA-HCTA and pSECRi-PFLA, whereas dh indicates E. coli DH5α cells with the pABA-HCTA and pSEVA221. Data represent the averages of three biological cultures, and error bars show the standard deviation (SD)