Project description:Despite decades of research, predicting where a transcription factor (TF) binds using its sequences alone continues to be an outstanding problem. Here, our thermodynamic model of Gal4p-DNA binding revealed a new mechanism of binding site selection in which Gal4p can bind to promoters that lacks it consensus site. Further investigation revealed that these sequences that Gal4p unexpectedly binds to have a high density of its half site, CGG. Analyses showed the ability of CGG half sites to predict zinc cluster TF in vivo binding and exposed a clear relationship between CGG half sites and zinc cluster binding in multiple binding datasets. To dissect the molecular logic of this half site mode of binding we designed a Sort-seq library containing 6,798 yeast promoter sequences and measured their expression. We learned that there is a linear relationship between CGG half site density and zinc cluster driven expression in vivo. Additionally, we found that promoters encoding 10 half sites have the same transcriptional output as a single consensus site and found no clear effects of half site orientation and spacing on strength of activation. Our experiments demonstrate that this novel mode of half site binding is widespread across the yeast genome and expands our knowledge about TF binding specificity that can be applied to higher eukaryotes.
Project description:Despite decades of research, predicting where a transcription factor (TF) binds using its sequences alone continues to be an outstanding problem. Here, our thermodynamic model of Gal4p-DNA binding revealed a new mechanism of binding site selection in which Gal4p can bind to promoters that lacks it consensus site. Further investigation revealed that these sequences that Gal4p unexpectedly binds to have a high density of its half site, CGG. Analyses showed the ability of CGG half sites to predict zinc cluster TF in vivo binding and exposed a clear relationship between CGG half sites and zinc cluster binding in multiple binding datasets. To dissect the molecular logic of this half site mode of binding we designed a Sort-seq library containing 6,798 yeast promoter sequences and measured their expression. We learned that there is a linear relationship between CGG half site density and zinc cluster driven expression in vivo. Additionally, we found that promoters encoding 10 half sites have the same transcriptional output as a single consensus site and found no clear effects of half site orientation and spacing on strength of activation. Our experiments demonstrate that this novel mode of half site binding is widespread across the yeast genome and expands our knowledge about TF binding specificity that can be applied to higher eukaryotes.
