Wingdings or bust

Fantastic! What an honor. Please feel free to share any critiques or criticisms!

I'd join such a list.

For me, this reshapes how I think about regulation. Every TF we look at has this useful behavior of buffering physiological and/or genetic perturbations to promoter activity. Although we expect, from basic models, many TFs should actively amplify these perturbations, they don't. Here's the link:

There's more in the paper, but here is an overarching figure. All the data from the paper (and some existing data) collapsed to the same basic, universal behavior. We never see evidence of the relationships that would come from "destabilizing" type interactions. Everything looks stabilizing.

We look at natural promoters. SoxS regulates three promoters (PoxB, FldA and DecR) at very similar positions. The same relationship is visible here covering all the way from 100-fold activation to 2-fold repression, all predictable from the basal activity of the promoter being regulated.

We also made a toy system using sigma28 promoters and controlling sigma28 concentration to change promoter activity. Really it's the same story over again. We can change promoter sequence, we can change active polymerase concentration and we see the same signatures of stabilization.

There's other ways to change promoter activity. We alter physiology with media: changing the growth rate changes the activity of each promoter. Here's the LacI and CpxR library. We find the same relationship. The slope across genetics (promoters) and/or physiology (different media) is -1.

Necessarily, what I mentioned previously is also true. Different promoters with ~100s fold differences in basal expression all produce the same level of regulated expression and that level is different for different TFs. This is very... useful? The expression of each promoter is set by the TF.

Here's five repressors: LacI, AscG, AcrR, MngR, PdhR. All of them have the relationship expected from a stabilizing TF, including LacI. This is surprising, especially given the standard textbook-level picture of these regulators as steric hindrance machines, interfering with RNAP binding (destab).

Here's three activators: CpxR, MetR, SoxS. They all have the expected relationship for stabilizing TFs. It's probably fair to say this is expected. As previously mentioned, activators often work through a mechanism involving a positive interaction between TF and RNAP at the promoter.

Our goal was to test this systematically. We used our Titration library (PMC9189660) to measure individual TFs regulating ~100 mutant promoters, measuring both unregulated expression and the maximum regulated expression of each promoter. We plot fold-change vs constitutive exp of each promoter.

Another prediction: the slope for stabilizers is always the same for strong promoters, -1. Here, changing constitutive levels wont change regulated levels. The TF buffers constitutive expression changes of the promoter. However, destabilizing TFs should amplify differences of similar promoters.

Our model predicts a relationship between promoter strength & regulation that depends only on the nature of a TFs stabilizing interactions: Stabilizers will exert lower fold-change on stronger promoters. Destabilizers, the opposite. So CpxR is a stabilizer? We previously inferred that (PMC8667592)

Stabilization is associated with activators making beneficial contacts with a domain on the RNAP. Destabilization is associated with repressors and "steric hindrance", i.e. exclusionary binding of TF that prevents RNAP from binding partly or fully. Example figures from Lloyd et al PMID: 11758454.

Let's focus quickly on "stabilization". This is the degree to which the TF promotes (stabilizes) or interferes with (destabilizes) the occupancy of RNAP at the promoter. Here we quantify it with the parameter beta, which is really just a measure of how much the TF impacts the TF-RNAP cobound state.

It is also helpful to ask "What were we expecting?" We use a simple model of regulation by two distinct mechanisms: the first (stabilization/destabilization) is the interaction between TF and RNAP at the promoter, the second is acceleration/deceleration of RNAP initiation from cobound TF-RNAP.

We tend to excuse these differences as "context dependent" but here context is largely controlled. Is this "context" or rather just a natural characteristic of TF function? Looking at the three promoters above, the two "weak" promoters were activated by CpxR and the "strong" one repressed.

HI Erik, would you mind adding me? My group is in the Department of systems biology at Umass Chan Medical School. We work on transcription regulation in bacteria. My group page is here: www.brewsterlab.net and my scholar profile is here scholar.google.com/citations?us.... Thanks!