October 20, 2017
Imagine that Jane is in trouble on the edge of the jungle. She needs to be saved soon or she will be sent back to Europe (which she does not want).
Tarzan knows he can’t get there in time by running along the jungle floor. But of course he has a trick up his sleeve—swinging from vine to vine!
He gets there in time, fends off her would-be kidnappers, and saves the day. All because he transferred from one strand of vine to another instead of “sliding” along the jungle floor.
A new study by Wollman and coworkers shows that at least two transcription factors in Saccharomyces cerevisiae, Mig1p and Msn2p, seem to share a lot in common with Tarzan. In a process termed intersegment transfer, they get to where they need to in the genome by “swinging” from DNA segment to DNA segment instead of just sliding along the DNA.
And transcription factors like these need to get to the right place in time to save the cell from outside threats. Just like Tarzan had to swing from vine to vine to save Jane.
This sort of approach would not necessarily work well with just a single transcription factor with a single DNA binding domain. It would sort of be like a one-armed Tarzan—it is hard to take advantage of swinging from vine to vine without at least two arms!
The authors argue that Mig1p gains its “extra arms” through joining together into a cluster. Now each Mig1p can bind DNA and drag the other transcription factors with them. A neat solution to the one arm problem.
Their basic approach is to use fluorescence microscopy to follow a GFP-Mig1p fusion protein in a single cell, something that has only become possible recently. What they found was that there were two populations of transcription factors—a diffuse set of smaller molecules and distinct, larger clusters made up of multiple Mig1p’s.
The clusters appeared to be the ones doing the work in the nucleus. Kinetic studies showed that they stayed in one place in the nucleus for over 100 seconds which is consistent with the clusters and not the monomers being bound to the DNA.
OK so this transcription factor tends to clump up into clusters and it looks like these clusters are the ones regulating gene expression. They also showed that a second transcription factor, Msn2p, did the same thing.
The authors next set out to see if this approach made sense for genetic regulation by running simulations of Mig1p finding its sites in the nucleus as either monomers or as clusters. It made sense to form clusters.
A lot of previous work has been done in S. cerevisiae in terms of the three dimensional map of the genome and where Mig1p DNA binding sites were located in this mesh of DNA. And in the course of their studies, Wollman and coworkers were able to estimate how many Mig1p molecules were in yeast cells and how many were in clusters.
They now had all the information they needed to run their simulations. When they crunched the numbers, they found that clusters fit their data much better than monomers (R2 = 0.75 vs. R2 < 0).
The final step was to work out what part of Mig1p was involved in forming the clusters. To do this, the authors compared Mig1p and Msn2p, the second transcription factor they studied that also formed into clusters, and looked for structural regions they might have in common.
What they found was both proteins had a highly disordered region. For Mig1p, it was at the C-terminus and for Msn2p it was at the N-terminus.
The hypothesis is that these disordered regions, which are both at the opposite end of the protein from the DNA binding domain, interact and form ordered structures that enable clusters to form. Wollman and coworkers used circular dichroism to show that when Mig1p was put in conditions that favor cluster formation, there was a transition consistent with unstructured protein becoming structured.
What we seem to have is a cluster of transcription factors connected in the middle with their DNA binding domains pointing out. This rolling cluster can more easily hop from DNA strand to DNA strand to find the right spots to bind.
Without this mechanism, Mig1p couldn’t get to where it needs to in time. It is as if Tarzan had 6-9 arms circling his body so he could get to Jane even more quickly.
With the wide range of tools available and our deep understanding of how yeast works and how a yeast cell is organized, our trusted ally S. cerevisiae again teaches us something fundamental about how our biology works.
Mig1p may be able to swing like Tarzan but it can’t yell like him. Or like Carol Burnett!
by Barry Starr, Ph.D., Director of Outreach Activities, Stanford Department of Genetics
Categories: Research Spotlight
April 24, 2012
A new study by Lickwar and coworkers suggests that many transcription factors fidget on and off the DNA, waiting for some signal to get to work. Once they get that signal, they clamp down and start affecting the activity of nearby genes.
If true this would help explain some perplexing results researchers have been getting with chromosomal immunoprecipitation (ChIP) assays. Transcription factors appear to be bound at many places where they are not affecting any nearby genes. Now we might have an idea why.
These researchers came up with this model through the use of an elegant, in vivo competition study. What they did was to set up a yeast strain that contained two different versions of the transcription factor Rap1p. One version was tagged with a FLAG epitope and was under the control of RAP1’s endogenous, constitutive promoter. The other version was tagged with a Myc epitope and was under the control of an inducible promoter.
They started out seeing where Rap1p was bound in the absence of the inducer by using an antibody against FLAG. This is the equivalent of a typical ChIP experiment. They found Rap1p was bound in many places throughout the genome including sites where it did not appear to affect any nearby genes.
Then they added the inducer galactose and at various time points repeated the ChIP experiment with antibodies against either FLAG or Myc. They were basically looking for how quickly the Myc-tagged Rap1p replaced the FLAG-tagged Rap1p with the idea that less stably bound transcription factors would be replaced more quickly.
They indeed found that some sites were better able to withstand the onslaught of Myc-tagged Rap1p. And more importantly, that these sites were near genes most influenced by Rap1p. In other words it appears that the more stably bound the Rap1p, the bigger the effect it has on nearby genes.
They then went on to show that more stable binding correlated with lower nucleosome occupancy and stronger in vitro binding. From this data they propose a model where the level of the effect on transcription is the result of a competition between nucleosome and transcription factor binding. Stronger transcription factor binding keeps nucleosomes away so transcription can proceed.
They took the model one step further and proposed that transcription factors are idling on the DNA, waiting for a signal to bind more tightly and influence the activity of nearby genes. In other words, transcription factors are ready to have an effect at a moment’s notice.
This part of the model still has to be proven though. All that has been shown so far is that a slow off rate is required for effective transcription activation by Rap1p. What we don’t know is whether this translates to other transcription factors or if idling Rap1p is ever more stably bound.
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
Categories: Research Spotlight