Preview

PDF, English - main document Download (13MB) | Terms of use

Abstract

The formation of multi-protein complexes is a key feature of the cellular proteome in all kingdoms of life. The biogenesis of protein complexes in vivo is still poorly understood, but recent methodological advances now make it possible to reveal the underlying mechanisms. One milestone method, termed Selective Ribosome Profiling (SeRP), allowed to demonstrate the omnipresence of co-translational assembly in bacteria and yeast (Shieh et al. 2015, Shiber et al. 2018). This method provides codon-resolved information about heterodimer formation between already completed and folded proteins and their nascent partner subunits (termed co-post assembly). The fact that assembly can occur during translation raised the question whether co-translational assembly could also involve interaction between two nascent proteins translated by two neighboring ribosomes (termed co-co assembly). So far, indirect evidence suggested co-co assembly of only a few protein complexes. However, direct evidence that two ribosome-nascent chain complexes interact via their nascent chains is still scarce and we lack any information about the prevalence of this proposed process.

This dissertation focused on the investigation of the hypothesized co-co assembly mode. In collaboration with Matilde Bertolini (PhD student in the Bukau lab), we first developed an unbiased, proteome-wide screen based on ribosome profiling (Disome Selective Profiling, DiSP), to reveal the prevalence of co-co assembly in human cells. By applying DiSP to HEK293-T and U2OS cells, we identified hundreds of high confidence co-co assembling nascent proteins. Our proteome-wide data suggest that up to 30% of all annotated homomer subunits employ co-co assembly, most frequently induced by the formation of N-terminal coiled coils (mostly partially exposed at assembly onset) or interactions of well-known globular dimerization domains (that are generally fully exposed at assembly onset). We further show that co-co assembly of two human homodimeric candidates can be recapitulated in bacteria, in the absence of any eukaryote specific machinery. This suggests that assembly is solely facilitated by the intrinsic propensities of the nascent proteins to form quaternary structures. In addition, we validate the existence of co-co assembly also for endogenous E. coli proteins by DiSP.

The main outcome of this dissertation is the demonstration of co-co assembly as a mechanism mainly employed for homomer formation. Our experimental data indicate the high prevalence of co-co assembly in human cells to ensure productive protein complex biogenesis in the crowded cytosol of cells. Initial findings suggest that co-co assembly is coordinated by a general transient slow-down of translation at the onset of assembly.