Context and number of non‐canonical repeat variable diresidues (RVDs) impede the design of TALE proteins with improved DNA targeting

A particularly promising approach for engineering more precise TALE proteins is to better understand non-canonical RVDs. The RVD repertoire employed by Xanthomonas extends well beyond just the four canonical RVD types, and natural TALE repeat arrays often utilize serial combinations of both canonical and non-canonical RVDs in a manner which is not fully understood. Attempts to characterize these non-canonical RVDs have utilized various approaches, in terms of both TALE design and experimental method. Some studies have considered just TALEs of particular biological relevance, others have looked at large numbers of different RVDs within a fixed repeat-array-context, and still others have designed custom TALEs to address hypotheses about optimal RVD positioning within the repeat array. Methods for inferring the DNA binding activity of these TALEs have included sequencing of cleaved DNA fragments from TALE-nuclease fusions; ELISA assays for TALE-oligonucleotide binding; reporter assays for TALEs’ transcriptional-activating effects; and in silico modeling. In this study we employ a novel “wholesale swap-out” approach to characterize 3 non-canonical RVDs for thymine and 4 non-canonical RVDs for guanine. In this study we employ a novel “wholesale swap-out” approach to characterize 3 non-canonical RVDs for thymine and 4 non-canonical RVDs for guanine. We assemble a library of 46 TALE proteins, comprising 11 reference TALEs bearing exclusively canonical RVDs, and 54 variant TALEs for characterizing the 7 non-canonical RVDs of interest. We design custom protein-binding microarrays (PBMs) which bear probes for the predicted target sequence of each TALE, in addition to probes for all possible mono- and di-nucleotide substitutions of these target sequences. The resulting quantitative binding data for the TALE proteins to ~5,000 unique DNA probes allows us to infer the relative affinity and specificity of each non-canonical RVD to each of the four nucleotides.