On January 26, Rob Brewster, a postdoctoral scholar from the California Institute of Technology, gave a seminar on his research on modeling gene expression from a biophysics perspective. His research focuses on using physical laws to create models that can predict gene expression. Any model inherently overlooks certain factors, which Brewster calls “inconveniences.” In order to account for the numerous inconveniences in gene expression, he hopes to numerically describe the “input and output” functions of gene expression in order to write the “Ohm’s law” of transcription.
The background images are cells with varying levels of repression of the green fluorescence gene. The foreground features the graph of green fluorescent protein created v. level of repression.
Source: Division of Biology and Biological Engineering, Caltech
Brewster’s lab tests his models in E. coli bacteria by using genetic and molecular biology techniques to finely adjust various characteristics of the many enzymes involved in gene expression. One example is altering the binding strength of RNA polymerase, the enzyme that creates the messenger RNA (mRNA) that eventually is translated into gene products. Two major current models of mRNA production, called transcription, are the thermodynamic model and the kinetic model. The thermodynamic model attempts to quantify the probabilities of the various configurations of RNA polymerase and other important enzymes around the cell. On the other hand, the kinetic model enumerates the many ways gene expression can progress, accounting for factors such as the rates of mRNA formation and degradation. Brewster’s work incorporates both of these models to predict transcription. After growing E. coli, transcription is measured using assays that quantify the amount of RNA and target protein present in cells.
One of the major inconveniences that Brewster and other members of the lab study is the sharing of limited amounts of regulatory proteins between genes, which produces competition between genes even though they may not directly interact. Using principles of the thermodynamic model, Brewster can calculate the probability of finding the RNA polymerase appropriately bound to DNA, which is proportional to gene expression. As the lab moved from simple models of isolated genes to models of multiple gene copies and more complex competition between genes, the predicted level of transcription began to agree with the protein assays and video microscopy data used to measure it.
The second major inconvenience that Brewster discussed was the cell-to-cell variability in gene expression levels. Brewster adopted an evolutionary perspective to understand the reasons for this “noise” in expression. In one particular study, bacteria cells were grown in the presence of two sugars, one of which is more favored. Cells use the favored sugar until it runs out, at which time most cells switch their cellular machinery to using the other, less-favored sugar. However, a minority of cells did not switch their machinery, attempting to “wait out” the lack of the favored sugar. When the culture was treated with an antibiotic, all of the cells that switched their machinery to the second sugar died, leaving the cells who did not to repopulate the colony. This suggests a tradeoff between growth rates and safety (1). Brewster’s models also suggest that noise in expression can be predicted with a model incorporating two main factors: the binding affinity of RNA polymerase to DNA and the activity of repressor molecules that prevent RNA polymerase activity
Brewster’s goal as he continues his research is to quantitatively understand how the genome regulates gene expression. He also hopes to study the effects of cellular resource sharing in more complex systems such as genes with both repressors and activators, which are proteins that increase transcription rates.
Source:
1. Brewster RC, Weinert FM, Garcia HG, Song D, Rydenfelt M, Phillips R (2014) The transcription factor titration effect dictates level of gene expression. Cell, 156, 1312-1323. doi:10.1016/j.cell.2014.02.022