UC Berkeley

Drosophila melanogaster has been a key model organism in the study of animal development for many decades. Since the revolutionary Nobel Prize winning work of Christiane Nüsslein-Volhard and Eric Wieschaus in the 1970's, we have understood that developmental patterning proceeds in stages wherein morphogen gradients provide the information to specify segments that give rise to distinct physiological compartments in the organism. However, knowing the genes involved and the qualitative features of developmental processes is only the first step on the road to truly understanding animal development. Detailed quantitative studies enabled by recent technological advances may allow us to go a step further and develop a predictive understanding of development and perhaps even engineer new developmental processes. As the Nobel Prize winning physicist Richard Feynman said, "What I cannot create, I do not understand." This simple idea was the driving force behind most of the work presented in this thesis. Standing on the shoulders of the giants in developmental biology, I have sought to create simple synthetic gene regulatory regions in the early fly embryo that are amenable to quantitative theoretical dissection. In addition, I have pursued experiments to challenge the theoretical assumptions underlying these models, namely how binding of activators leads to transcriptional activation and also developed new tools to enable these theoretical studies.

We have developed a minimal synthetic enhancer containing a single Dorsal binding site for the Dorsal activator as a tool for theoretical dissection of transcriptional regulation in the early Drosophila melanogaster embryo. We found that a simple, theoretical model of transcriptional dynamics is sufficient to explain the fraction of loci in the embryo that engage in transcription and the timing of their transcription.

I also investigated transcription by Bicoid driven minimal synthetic enhancers. In contrast to the Dorsal activator, a single Bicoid binding site was found to be less capable of specifying positional information in the embryo, perhaps due to its extremely rapid on rates, which we measured during our binding studies on the lattice light sheet. Nonetheless, we made headway in developing a synthetic platform for studying the Bicoid activator, chiefly by creating transcription factor 'neutral' reporter sequences which do not bind other early embryonic transcription factors. Further study will be required to push the Bicoid synthetic platform to investigate similar questions possible using the Dorsal activator.

In addition to studying transcription driven by the Bicoid activator, we studied its binding kinetics in living embryos in an effort to shine light on assumptions about the binding events that precede transcriptional activation. We have pushed the envelope of the in vivo imaging of developing organisms by using the lattice light sheet microscope to probe the binding kinetics of single Bicoid molecules to the fly genome. We discovered that, contrary to expectations, Bicoid tends to bind in spatially localized clusters in the nucleus, and that these clusters facilitate binding to low-affinity clusters in the posterior embryo. Furthermore, the transcriptional pioneering factor Zelda is necessary for this clustering behavior and potentiates binding along the full length of the anteroposterior axis.

Finally, we developed tools for quantitative studies in the early fly embryo. During the course of many other projects, including the ones outlined above, we have found that precise levels of maternally deposited proteins in the early embryo are often critical to quantitative studies, but there is a dearth of 'tunable' maternal promoter sequences available with well-characterized behaviors. We thus sought to develop such promoters, leading to the creation and characterization of promoters capable of driving the expression of maternally deposited proteins spanning several orders of magnitude in concentration. We additionally explored the effects of genomic location on the expression levels of our modular maternal promoters expression levels, which should expand their utility in the future. Lastly, we developed methods for calibrating measured fluorescence values from microscopy images into absolute numbers of molecules, an invaluable tool in quantitative studies.

To summarize, the experiments presented in this thesis lay the foundations for studying transcription in development with high spatiotemporal resolution and quantitative precision, contributing to a thoroughly predictive understanding of developmental processes.