Protein-DNA and protein-protein interactions of polycomb group transcriptional repressors from Drosophila melanogaster

Polycomb group (PcG) repressor proteins maintain the spatially restricted expression of homeotic genes during animal development. Biochemical and genetic data support a model where several distinct multimeric complexes act in concert at the chromatin of target genes to block transcription. The studies presented in this thesis explore the molecular properties of PcG proteins from Drosophila melanogaster to better understand the mechanism of PcG repression. PHO, the only known DNA-binding PcG protein, was tested for binding to regulatory regions from homeotic genes. Several binding sites were found in two different elements that correspond to the sequence GCCAT. To test the possibility that PHO recruits other PcG proteins to regulatory elements, a panel of PcG proteins was tested for direct interaction with PHO. No interaction was observed, prompting the suggestion that PHO may instead play a supporting role in PcG function. Two other PcG proteins, SCM and PH, contain an SPM domain. In vitro binding experiments define the SPM domain as a multimerization motif. A collection of mutations was identified in the SPM domain of SCM that affect its binding properties. Several of these mutant domains were tested in a Scm rescue assay. Domains with reduced binding affinity did not support Scm function, proving that the SPM domain is a critical region of the SCM protein. A specificity mutant that retains PH but not SCM binding also failed to rescue in vivo, indicating that SCM-PH binding is not sufficient for function. Biochemical analysis revealed that this mutant SCM protein assembled into a 500 kDa complex along with the majority of endogenous SCM. This result suggests that the SPM domain is required not for complex assembly, but rather for chromosomal localization or for interactions between PcG complexes at the site of repression. Overexpression of an isolated SPM domain was lethal when widely expressed during development, and limited expression produced homeotic defects consistent with PcG loss-of-function. This work is important because it defines a protein interaction domain required for transcriptional repression during animal development, and also provides a set of tools for mechanistic studies of PcG repression.