Towards Highly Multigenic Manipulation of Gene Expression

Abstract

Biological processes rely on the coordinated expression of many genes to produce a phenotype. When and where in the body genes are expressed is tightly regulated on multiple, interconnected levels by the three-dimensional organization of the genome, post-translational modification of histone proteins, chromatin remodeling complexes, and transcription factor activity. Gene regulation is also a highly dynamic process, allowing cells to dramatically change their gene expression profiles in response to signaling molecules like hormones, or during processes like lineage specification. These gene regulatory processes are, in turn, governed by combinations of factors. For example, many genes are regulated by multiple cis-regulatory elements that can tune their expression in a context-dependent manner, and many cis-regulatory elements require binding by multiple transcription factors in order to modulate gene expression. However, methods that allow researchers to perturb many genes or regulatory elements at once, especially in high-throughput, are lacking. Since higher order interactions between factors can be difficult to predict from individual perturbations, such as in cases of functional buffering or synergy among genes or regulatory elements, this represents an important gap in our ability to model gene regulatory biology using experimental techniques. The work presented in this dissertation has two major goals: 1) to deepen our understanding of endogenous cellular processes responsible for facilitating rapid gene expression changes in response to a stimulus; and 2) to develop technologies and methods that enable researchers to perturb many genes or regulatory elements at once, including in high-throughput screens. To address the first goal, we performed a genome-wide screen to identify factors that are necessary for the glucocorticoid-induced expression of the protein GILZ and found that the chromatin remodeling proteins SMARCA2 and BPTF played a role in glucocorticoid-induced GILZ expression. The glucocorticoid response is an excellent model for studying multigenic gene regulation since > 1,000 genes undergo a significant change in expression after glucocorticoid treatment. I interrogated the genome-wide impact of SMARCA2 and BPTF knockdown on glucocorticoid-induced changes in gene expression and chromatin accessibility and found that BPTF plays a very limited and specific role in the glucocorticoid response while SMARCA2 has a much broader role, with an estimated 30% of glucocorticoid-responsive genes having an altered response after SMARCA2 knockdown. I found that SMARCA2 is necessary for glucocorticoid-induced increases in chromatin accessibility at a small fraction of differentially accessible regions of the genome. These SMARCA2 dependent regions are enriched near glucocorticoid-responsive genes, experience glucocorticoid-induced binding of many transcription factors, and have high glucocorticoid-induced regulatory activity as measured by STARR-seq. Taken together, these data suggest that SMARCA2 contributes to changes in the expression of many genes after glucocorticoid treatment by facilitating an increase in accessibility at key regulatory elements. To address the second goal, I used CRISPR-Cas12a epigenome editing technology to manipulate the expression of many genes at once, and performed two multiplexed, high-throughput screens. I validated two novel dCas12a epigenome editors, and demonstrated that dCas12a epigenome editing can be successfully employed in cultured primary human T cells and be used to differentiate iPSCs into neurons. I also developed a system for simultaneous activation and repression of defined target genes using orthogonal Cas12a proteins. To demonstrate the potential of dCas12a for high-throughput screening, I performed a fitness screen using four dCas12a repressors and showed that the dHyperLbCas12a variant is suitable for use in screens while the dEnAsCas12a variant is not. Lastly, I performed a high-throughput screen to determine how two enhancer elements at the PER1 locus interact to control gene expression in response to different doses of glucocorticoids.