High-throughput functional characterization of combinations of transcriptional activators and repressors

Despite growing knowledge of the functions of individual human transcriptional effector domains, much less is understood about how multiple effector domains within the same protein combine to regulate gene expression. Here, we measure transcriptional activity for 8,400 effector domain combinations by recruiting them to reporter genes in human cells. In our assay, weak and moderate activation domains synergize to drive strong gene expression, while combining strong activators often results in weaker activation. In contrast, repressors combine linearly and produce full gene silencing, and repressor domains often overpower activation domains. We use this information to build a synthetic transcription factor whose function can be tuned between repression and activation independent of recruitment to target genes by using a small molecule drug. Altogether, we outline the basic principles of how effector domains combine to regulate gene expression and demonstrate their value in building precise and flexible synthetic biology tools. Demultiplexed sequencing reads were provided by Admera. Domain 1 and domain 2 reads were aligned separately using bowtie2, trimming 30 base pairs from both 5’ and 3’ ends, and were then each paired read was identified as a particular library member using a python script iterating over each read pair. The total number of reads for each library member was then summed up using another python script. Library members were required to have at least 5 ON reads or 5 OFF reads as well as 50 total reads across both ON and OFF subpopulations to be considered for downstream analysis. For every qualifying library member, the ON or OFF read count was then set to 5 if it was less than 5. Then, ON read counts were normalized by dividing each library member’s read count by the total number of ON reads, and the same was done for OFF counts. Overall enrichment scores were then computed by taking the log2 of the ratio of normalized ON read counts to normalized OFF read counts.