In this paper, we examined the relative contributions of gene transcription and protein translation to the cost of protein production in budding yeast, S. cerevisiae, and further mapped the limitation to the initiation versus elongation steps of each process.
Protein production in living cells is tightly coordinated with external conditions and intracellular demands. This regulation ensures that needed proteins are produced, while those whose function is not compatible with current needs are not. It may also serve to minimize the cost of protein production; indeed, making proteins consumes cellular resources by using energy and nutrients as building blocks and also by occupying common cellular machineries such as ribosomes, polymerases, or chaperones, whose abundance may be limiting.
TDH3 promoter driving mCherry expression (pTDH3-mCherry) was cloned into a plasmid that, when transformed into yeast cells, integrates into the genome in multiple copies (left). mCherry expression range can be appreciated from the differences in the colonies’ color (right). Fluorescence images of individual cells expressing low or high mCherry copies and superimposed with bright light (right).
Quantifying growth fitness: Cells expressing a given level of mCherry were co-incubated with a GFP-labeled wild-type strain for 30 generations, during which, their relative frequencies were quantified using a flow cytometer (left). Relative growth rate defined with respect to wild-type cells was calculated from the rate by which the frequency of the mCherry expressing cells was reduced (middle). The fitness cost of protein expression: The relative growth rates of mCherry-expressing cells as a function of mCherry fluorescence (right).
Genomic integrations do not contribute significantly to the measured fitness cost, note the negligible reduction in growth rate for the DAmP-RPS5-based, which mean minor fitness cost. Protein toxicity does not contribute significantly to the measured fitness cost. Destabilized version of GFP (open gray, CLN2 degradation box) shows the same fitness cost as the GFP strains, which mean that no extra or less cost is observed in these strains.
DAmP - Decreased Abundance by mRNA Perturbation (Muhlrad and Parker, 1999; Schuldiner et al., 2005).
Mapping the protein production limitation in the different conditions: Low Pi and Low N mapped almost exclusively to transcription initiation and to translation elongation, respectively. Further, in SC condition the cost originate both from transcription initiation and translation initiation.
Low and high haploid burdened strains were mated with a library of 90 GFP-fused genes and then subjected to SGA protocol, resulting in new haploids containing both the GFP-fused gene and our mCherry burden. The low- and high-burden strains for each GFP gene (left and right density clouds, respectively) grew together in the same well, allowing a highly controlled comparison of the GFP levels. Black dashed lines represent the GFP and mCherry medians of the two strains.
For each mCherry copy-numbers bin, the median of GFP fluorescence ratio (high burden/low burden) was calculated. The endogenous gene levels are shown with the binned mCherry copy number, and a linear fit was plotted (dashed line)

The Cost of Protein Production

By Moshe Kafri*, Eyal Metzl-Raz*, Ghil Jona, Naama Barkai

Link to Manscript


Questions: all,