Comparing DNA binding specificities between orthologous Pho4 transcription factors

Combinatorial control by multiple transcription factors (TFs) is a hallmark of eukaryotic gene regulation. Despite its prevalence and crucial roles in enhancing specificity and integrating information, the mechanisms behind why eukaryotic TFs depend on one another, and whether such interdependence evolves, are not well understood. We exploit natural variation in co-TF dependence in the yeast phosphate starvation (PHO) response to address this question. In the model yeast Saccharomyces cerevisiae, the main TF, Pho4, relies on the co-TF Pho2 to regulate ~28 genes. In a related yeast pathogen, Candida glabrata, its Pho4 exhibits significantly reduced Pho2 dependence and has an expanded target set of ~70 genes. Chomatin IP based motifs for the two orthologous Pho4s are highly similar, with no detectable difference in their core E-box motif. However, past studies found that preference for sequences flanking the core motif can provide additional specificities to TFs in the same family. To determine if ScPho4 and CgPho4 differ in their flanking nucleotide preference, we conducted the universal Protein Binding Microarray (uPBM) to interrogate all possible 9-mers, from which a robust 7-mer enrichmement score can be calculated. Universal protein-binding microarray (PBM) experiments were performed for recombinant, full-length and orthologous transcription factors Pho4 from two yeast species, S. cerevisiae and C. glabrata. Briefly, universal PBMs involved binding of GST-tagged transcription factors to double-stranded 8x15K Agilent microarrays containing a DNA library designed to cover all possible 9-bp sequences, with every 8-mer occurring in at least 16 different spots on the array. This design allows comprehensive and unbiased characterization of the binding specificity of transcription factors for all possible 7-bp sequences.