Phenotypic engineering by reprogramming gene transcription using novel artificial transcription factors in Escherichia coli

Ju Young Lee 2 Bong Hyun Sung 2 Byung Jo Yu 2 Jun Hyoung Lee 2 Sang Hee Lee 2 Mi Sun Kim 1 Michael D. Koob 0 Sun Chang Kim 2 0 Department of Laboratory Medicine and Pathology, Institute of Human Genetics, University of Minnesota , Minneapolis, MN 55455, USA 1 Biomass Team, Korea Institute of Energy Research , Daejeon, Korea 2 Department of Biological Sciences, Korea Advanced Institute of Science and Technology , Daejeon Now that many genomes have been sequenced and the products of newly identified genes have been annotated, the next goal is to engineer the desired phenotypes in organisms of interest. For the phenotypic engineering of microorganisms, we have developed novel artificial transcription factors (ATFs) capable of reprogramming innate gene expression circuits in Escherichia coli. These ATFs are composed of zinc finger (ZF) DNA-binding proteins, with distinct specificities, fused to an E. coli cyclic AMP receptor protein (CRP). By randomly assembling 40 different types of ZFs, we have constructed more than 6.4 104 ATFs that consist of 3 ZF DNA-binding domains and a CRP effector domain. Using these ATFs, we induced various phenotypic changes in E. coli and selected for industrially important traits, such as resistance to heat shock, osmotic pressure and cold shock. Genes associated with the heat-shock resistance phenotype were then characterized. These results and the general applicability of this platform clearly indicate that novel ATFs are powerful tools for the phenotypic engineering of microorganisms and can facilitate microbial functional genomic studies. - INTRODUCTION Historically, many unique cellular traits of microorganisms have been identified and modified, not only for the benefit of human life, but also for industrial applications. To improve the characteristics of useful microorganisms, several approaches have been explored to expedite the screening for new phenotypes of interest. Because of its simplicity and convenience, random mutagenesis has been widely used for the selection of desired phenotypes (1,2). However, mutagenesis approaches rely on labor-intensive work and luck. Therefore, more rational approaches have been employed for the manipulation of genes relevant to a specific function (35). Most rational approaches involve either the deletion or overexpression of a single gene or sequential multigenic modifications (4,5). However, the engineering of desired phenotypes in organisms often requires the complete reprogramming of innate gene expression circuits, a process that necessitates multigenic transcriptional coordination. Even with the most sophisticated computational methods for physiological pathway analysis, complete identification of the precise genes involved in achieving a desired phenotype is almost impossible. Therefore, to permit multiple simultaneous gene expression changes, researchers recently developed methods for engineering of the global transcription machinery (6,7). Although this methodology permits changes in the expression of many genes that allow organisms to access novel cellular phenotypes, its use has been restricted either to a specific genotype or to housekeeping genes that are generally expressed and thought to be involved in routine cellular metabolism (7). Here, we have developed novel artificial transcription factors (ATFs) to completely reprogram innate gene expression circuits in Escherichia coli and thus to elicit broad perturbations in the E. coli transcriptome. Zinc fingers (ZFs) are well-characterized and highly specific DNA-binding domains found in a wide variety of transcriptional regulatory proteins (8,9). Because of their diversity and modular structure, ZF domains have provided an attractive framework with which to construct diverse, novel ATFs (1017). The cyclic AMP receptor protein (CRP) is a global transcription factor that regulates gene expression at >200 different promoters in E. coli (1821). CRP activates transcription by interacting with RNA polymerase through CRPs functionally independent transcriptional activation domains. Because of the existence of these various domains (20), CRP could be engineered and used as a transcriptional effector domain to construct novel ATFs. In this study, we selected 40 human ZFs with diverse DNA-binding specificities and combinatorially assembled them to construct a library of DNA-binding domains that consisted of 3 ZFs each. These DNA-binding domains were then fused to a CRP effector domain to create a fusion protein that either activated or repressed transcription regardless of the presence of endogenous transcription factors. Introduction of these novel ATFs into E. coli induced various phenotypic changes such as resistance to heat shock, osmotic pressure and cold shock. The bacterial strains were selected that displayed the newly acquired industrially important traits and the genes associated with the selected phenotypes were identified and characterized. MATERIALS AND METHODS Strains and plasmids E. coli XL1-Blue (Stratagene, La Jolla, CA) and E. coli K-12 MG1655 were used for this experiment. To test an EGFP-based reporter system, MG1655DaraBAD was constructed by -Red-mediated markerless deletion (22). For complementation assays, the genes related to the thermotolerance phenotype, cpxP, ompW and the marRAB operon were deleted from the MG1655 E. coli genome by the above method, producing MG1655DcpxP, MG1655DompW and MG1655DmarRAB, respectively. The T7lac promoter of the pETDuet-1 vector (Novagen, San Diego, CA) was replaced by either a tac promoter or an arabinose-inducible promoter, generating pETtac and pETara, respectively. The pBR322 origin of the high-copy plasmid pETtac was replaced by P15A origin, generating low-copy plasmid pACYCtac. pETtac was used for the expression of the ATF libraries. pACYCtac was used for the expression of the T2 ATF to assess effects of T2 ATF copy number on the thermotolerance phenotype. In addition, for the complementary assays, the genes related to the thermotolerance phenotype, cpxP and ompW, were amplified with the polymerase chain reaction (PCR) from the MG1655 genomic DNA, and the amplified cpxP and ompW genes were cloned into pETtac individually or together, producing pETtac-cpxP, pETtac-ompW and pETtac-cpxP/ompW, respectively. pETara was used for the expression of the 3-ZF-CRP fusion proteins, which consisted of 3 ZFs fused to various CRP derivatives, to find the most effective CRP effector domain for this study (Figure 1A and Supplementary Table 2). Selection of a potent CRP effector domain with an EGFP-based reporter system Reporter plasmids (pEGFP-A and pEGFP-R) were constructed by inserting a tac promoter sequence containing the target DNA-binding site for the 3-ZF-CRP fusion proteins and the gene encoding enhanced green fluorescent protein (EGFP) into pACYC184 (New England Biolabs, Beverly, MA). pEGFP-A was constructed by inserting 2 copies of the target DNA sequence (50-GCG GCG GGG-30) upstream of (...truncated)