Despite the availability of whole-genome sequences for almost all model organisms, making faithful predictions of gene expression levels based solely on the corresponding promoter sequences remains a challenge. Even with advances in understanding of promoter structure and function, the contribution of the vast majority of base pairs to promoter architecture is currently unknown. The yeast GAL 1/GAL10 bidirectional promoter is arguably the best-studied promoter in eukarwtic biology. however information about its regulatory structure is limited to the classification of a few binding sites. This problem is largely due to limitations in the ability to assess the impact of mutations in vivo. Current methods such as plasmid-based approaches and integration using selection markers are not ideal due their disruptive nature and tendency to cause copy-number fluctuations. Here I present the development of a genome editing method utilizing the CRISPR/Cas9 complex which overcomes these problems in live yeast cells. The method involves the introduction of a novel cut site into a specific genomic location, followed by the integration of an edited sequence into the same location in a scarless manner. I then used this method to edit the GAL1 and GAL8O promoter sequences, and found that the relative positioning of promoter elements is critically important for the determination of promoter activity levels in single cells. Following this, I identified new regulatory regions within the GAL1/GAL10 promoter which control the direction of expression and which were unpredicted by current models of promoter activation. I further show that two sequences previously thought to be Gal4 bindingsites are actually associated with the newly discovered regulatory regions. Investigation of the PHO5 promoter reveals that the new regulatory mechanism is not specific to the GAL1 promoter and may be widespread in the genome. The discovery of new and unpredicted regulatory elements in this well-characterized promoter implies that eukaryotic promoter structure may be considerably more complex than previously thought.