Missense variants in human forkhead transcription factors reveal determinants of forkhead DNA bispecificity

Recognition of specific DNA sequences by transcription factors (TFs) is a key step in transcriptional control of gene expression. While most forkhead (FH) TFs bind either an FKH (RYAAAYA) or an FHL (GACGC) recognition motif, some FHs can bind both motifs. Mechanisms that control whether a FH is monospecific versus bispecific have remained unknown. Screening a library of 12 reference FH proteins, 61 naturally occurring missense variants including clinical variants, and 22 designed mutant FHs for DNA binding activity using universal (“all 10-mer”) protein binding microarrays (PBMs) revealed non-DNA-contacting residues that control mono- versus bispecificity. Variation in non-DNA-contacting amino acid residues of TFs is associated with human traits and may play a role in the evolution of TF DNA binding activities and gene regulatory networks. Variants of human forkhead gene DNA-binding domains were generated by site-directed mutagenesis of synthesized reference alleles and Gateway cloned into pDEST15. The resulting GST fusion proteins were expressed by in vitro transcription and translation. Agilent microarrays (AMADID #030236) containing all 10-nt sequences were made double-stranded by primer annealing and on-array extension with Cy3-labelled nucleotides to monitor double-stranding efficiency, then WT and variant GST-tagged proteins were applied and detected with Alexa488-labelled anti-GST antibody. Alexa488 signal was scanned at multiple intensities; scans were combined with Masliner and the resulting file was normalized to Cy3 signal and analyzed with the Universal PBM Analysis Suite (Berger et al., Nature Biotechnology 2006) to report Enrichment scores for each 8-mer.