Assembly of designer TAL effectors by Golden Gate cloning

Abstract

Generation of customized DNA binding domains targeting unique sequences in complex genomes is crucial for many biotechnological applications. The recently described DNA binding domain of the transcription activator-like effectors (TALEs) from Xanthomonas consists of a series of repeats arranged in tandem, each repeat binding a nucleotide of the target sequence. We present here a strategy for engineering of TALE proteins with novel DNA binding specificities based on the 17.5 repeat-containing AvrBs3 TALE as a scaffold. For each of the 17 full repeats, four module types were generated, each with a distinct base preference. Using this set of 68 repeat modules, recognition domains for any 17 nucleotide DNA target sequence of choice can be constructed by assembling selected modules in a defined linear order. Assembly is performed in two successive one-pot cloning steps using the Golden Gate cloning method that allows seamless fusion of multiple DNA fragments. Applying this strategy, we assembled designer TALEs with new target specificities and tested their function in vivo.

Figure 1. General overview of the two-step cloning strategy for dTALEs assembly.

(A) Golden Gate cloning principle applied for assembly of dTALEs. Plasmids encoding selected repeat modules (an example with only two modules, R1 and R2, is shown here due to space limitation) are mixed in one tube together with BsaI, T4 DNA ligase and the destination vector (containing a lacZα fragment for blue-white selection). Assembly of R1 and R2 using BsaI and ligase gives rise to a plasmid lacking the initial BsaI sites, but containing a block of assembled repeats flanked by two BpiI sites. The two BpiI sites allow release of the assembled repeats as one block for the second step of cloning. fs, fusion site. (B) Structure of AvrBs3. AvrBs3 contains a central region with 17 direct repeats (light grey boxes) flanked by a thymidine-specific repeat (repeat 0) and a half repeat (repeat 17.5, both flanking repeats shown as dark grey boxes). Two nuclear localization sequences (NLS, black bars) and a transcription activation domain (AD) are located in the C-terminal region. One representative 34 aa repeat is shown, with the RVD of the NI type highlighted in grey. (C) RVD types and their specificities. (D) Set of 68 repeat modules, with 4 modules with different specificities for each of the 17 repeat positions. Repeat modules are flanked by two BsaI sites with fusion sites selected from the codon-optimized sequence of AvrBs3 (see Supporting Information S1). Sets of five (for repeats 13–17) or six (for repeats 1–6 and 7–12) selected repeat modules are preassembled via BsaI into preassembly vectors (pL1-TA1 to 3). Preassembled repeat blocks are then combined in the final destination vector (pL2-TA) using a second BpiI-based Golden Gate cloning reaction. Construction of dTALE-1 is shown as an example.

Figure 2. Design and functional test of customized TAL effectors.

(A) Structure of the reporter construct present in transgenic N. benthamiana plants. The reporter construct contains a TMV-based viral vector construct under control of the alcA promoter. The vector contains the RNA-dependent RNA polymerase (RdRp) and a GFP gene, but lacks the viral movement and coat protein genes. Viral vector-mediated GFP expression is obtained only in cells where the alcA promoter has been activated. Sequences selected for engineering of dTALE-1 to dTALE-3 are indicated by a black line. The transcription start site of the TMV-based vector is marked by an arrow. (B) Schematic representation of dTALE-1 to 4 constructs. (C) Agrobacterium tumefaciens strains containing dTALE-1 to dTALE-4 constructs were inoculated into leaves of transgenic plants. An empty Agrobacterium strain was also inoculated as a negative control (neg). GFP expression was analyzed 5 days after inoculation under UV light. dTALE-1, 2 and 3, which target sequences in the alcA promoter, induced GFP expression. In contrast, dTALE-4, which targets a randomly selected sequence (not present in the promoter), did not induce any GFP expression.