Comprehensive interrogation of natural TALE DNA-binding modules and transcriptional repressor domains

Abstract

Transcription activator-like effectors are sequence-specific DNA-binding proteins that harbour modular, repetitive DNA-binding domains. Transcription activator-like effectors have enabled the creation of customizable designer transcriptional factors and sequence-specific nucleases for genome engineering. Here we report two improvements of the transcription activator-like effector toolbox for achieving efficient activation and repression of endogenous gene expression in mammalian cells. We show that the naturally occurring repeat-variable diresidue Asn-His (NH) has high biological activity and specificity for guanine, a highly prevalent base in mammalian genomes. We also report an effective transcription activator-like effector transcriptional repressor architecture for targeted inhibition of transcription in mammalian cells. These findings will improve the precision and effectiveness of genome engineering that can be achieved using transcription activator-like effectors.

Figure 1. Identification of an optimal guanine-specific repeat variable diresidue (RVD)

a, Design of the TALE RVD screening system. Each RVD screening TALE (RVD-TALE) contains 12.5 repeats with RVDs 5 and 6 substituted with the 23 naturally occurring RVDs, and is fused to a Gaussia luciferase gene via a 2A peptide linker. The truncations used for the TALE is marked at the N- and C- termini with numbers of amino acids retained (top). Four different base-specific reporters with A, T, G, and C substituted in the 6th and 7th nucleotides of the binding site are used to determine the base-specificity of each RVD (middle). Each reporter is constructed by placing the TALE binding site upstream of a minimal CMV promoter driving Cypridina luciferase (bottom). b, Base-preference of each natural RVD (top) is determined by measuring the levels of relative luminescence unit (RLU) for each base-specific reporter after background subtraction and normalization based on TALE protein expression level (top). We clustered RVDs according to their base-preference after performing one-way analysis of variance (ANOVA) tests on each RVD. For RVDs with a single statistically significant reporter activity (p < 0.05, one-way ANOVA), we plotted the reporter activity of the preferred base above the x-axis, whereas the reporter activities for the non-preferred bases are shown below the x-axis as negative. We clustered and ranked the RVDs without a single preferred base according to their total activity level. The abundance of each RVD in natural TALE sequences, as determined using all available Xanthomonas TALE sequences in GenBank, is plotted on a log scale (bottom). All bases in the TALE binding site are color-coded (purple for A, red for T, orange for G, and blue for C). NLS, nuclear localization signal; VP64, VP64 viral activation domain; 2A, 2A peptide linker; Gluc, Gaussia luciferase gene; minCMV, minimal CMV promoter; Cluc, Cypridina luciferase gene; polyA signal, poly-adenylation signal. All results are collected from three independent experiments in HEK 293FT cells. Error bars indicate s.e.m.; n = 3.

Figure 2. Characterization of guanine-specific repeat-variable diresidues (RVDs)

a, Specificity and activity of different guanine-targeting RVDs. Schematic showing the selection of two TALE binding sites within the CACNA1C locus of the human genome. The TALE RVDs are shown above the binding site sequences and yellow rectangles indicate positions of G-targeting RVDs (left). Four different TALEs using NN, NK, NH, and HN as the putative G-targeting RVD were synthesized for each target site. The specificity for each putative G-targeting RVD is assessed using luciferase reporter assay, by measuring the levels of reporter activation of the wild-type TALE binding site and mutant binding sites, with either 2, 4, or all guanines substituted by adenine. The mutated guanines and adenines are highlighted with orange and green respectively. b, Endogenous transcriptional modulation using TALEs containing putative G-specific RVDs. TALEs using NN, NK, NH, and HN as the G-targeting RVD were synthesized to target two distinct 18bp target sites in the human CACNA1C locus. Changes in mRNA are measured using qRT-PCR as described previously. VP64, VP64 transcription activation domain. All results are collected from three independent experiments in HEK 293FT cells. Error bars indicate s.e.m.; n = 3.

Figure 3. Computational analysis of TALE RVD specificity

We performed extensive free energy perturbation (FEP) calculations for the relative binding affinities between the TALE and its bound DNA. Images show the three-dimensional configuration and results of the free energy calculation for NN:G (a) and NH:G (b) interactions from one repeat in the TALE-DNA complex. The second amino acid of the guanine-recognizing RVD (i.e., asparagine for RVD NN and histidine for RVD NH) and the guanine base of the bound double-stranded DNA are presented in space filling model and labeled. The free energy calculation results are listed below their corresponding structures.

Figure 4. Development of aTALE transcriptional repressor architecture

a, Design of SOX2 TALE for TALE repressor screening. A TALE targeting a 14bp sequence within the SOX2 locus of the human genome was synthesized as described previously. b, List of all repressors screened and their host origin (left). Eight different candidate repressor domains were fused to the C-term of the SOX2 TALE. c, The fold decrease of endogenous SOX2 mRNA is measured using qRT-PCR by dividing the SOX2 mRNA levels in mock transfected cells by SOX2 mRNA levels in cells transfected with each candidate TALE repressor. d, Transcriptional repression of endogenous CACNA1C. TALEs using NN, NK, and NH as the G-targeting RVD were constructed to target a 18bp target site within the human CACNA1C locus (site 1 in Figure 2). Each TALE is fused to the SID repression domain. NLS, nuclear localization signal; KRAB, Krüppel-associated box; SID, mSin interaction domain. All results are collected from three independent experiments in HEK 293FT cells. Error bars indicate s.e.m.; n = 3. * p < 0.05, Student’s t test.