Researchers at the Stowers Institute for Medical Research have identified a number of protein complexes involved in the activation of endoplasmic reticulum stress response genes by the transcription factor ATF6α.
Detailed in a paper published last week in the Journal of Biological Chemistry, the study identified several proteins not previously known to operate in the context of ATF6α and offers an example of a sensitive, semi-quantitative mass spec-based workflow for profiling proteins involved in transcription, Joan Weliky Conaway, a Stowers researcher and author on the paper, told ProteoMonitor.
The research team, which included Weliky Conaway's husband Ronald Conaway – also a Stowers scientist – employed an LC-MS/MS workflow put together by their co-author Michael Washburn, director of the institute's proteomics center. Mass spec, Weliky Conaway said, has emerged as a key tool in studies of transcriptional regulation, offering significant gains in sensitivity and throughput over traditional chromatography-based methods.
"Traditionally there have been several approaches that [researchers] have used,” she said. "The first and certainly most labor intensive is assay-based chromatography, where you start with a biochemical assay where you're seeing good regulation of a particular promoter or particular gene in a test tube and then you work to purify [the elements present.]"
In addition to being labor intensive, such approaches also require a significant amount of sample, Weliky Conaway added, noting that using standard chromatographic techniques the researchers would have had to start with "kilograms of rat liver."
"Now, [with mass spec-based approaches], we can start with a few plates of tissue culture cells and have only a vanishingly small amount of enriched proteins to work with," she said.
In the JBC paper, the researchers set out to study the proteins recruited by ATF6α when it activates transcription of the ER stress response gene HSPA5, which has been implicated in signal transduction pathways linked to diseases including atherosclerosis, diabetes, and neurodegeneration. To do this, they immobilized on magnetic beads biotinylated DNA fragments containing a region of the HSPA5 gene that includes the ER stress response enhancer elements, core promoter, and early transcribed region. They then incubated these with two sets of HeLa cell extracts – one spiked with ATF6α and one without – and they eluted the bound proteins and analyzed them using a Thermo Fisher Scientific LTQ ion trap mass spec.
Using this workflow, the Stowers team identified a number of protein complexes including the Mediator RNA polymerase II coregulatory complex and the histone acetyltransferase complexes SAGA and ADAC that are likely involved in ATF6α function.
These complexes were largely known, but many hadn't been observed in the context of ATF6α, Weliky Conaway said, noting that the researchers had started with a relatively well-characterized genetic system with the notion of testing the workflow's effectiveness.
"Now that this approach has been shown effective, we'd like to extend it to genes about which there is less known, as well as genes that are regulated by substantially larger regulator regions and which might have more complex regulatory processes," she said.
As the researchers extend the technique to other genes, "there's a reasonable possibility that some of the co-activator or co-regulator proteins will be novel," Ronald Conaway told ProteoMonitor.
And "even if one isn't identifying novel co-regulators, this approach has made it possible in an unbiased way to ask what the co-regulators are that work with" a given transcription factor, Weliky Conaway said. "The analysis of [ATF6α] has raised a lot of questions about how these co-regulators function together, and so there's a lot of mechanistic work that we would like to pursue."
They also hope to perform similar experiments using DNA assembled in a "chromatin milieu" more representative of an actual in vivo system, she said.
One of the more significant challenges in the work was separating proteins involved with the gene and transcription factor being studied from non-specific DNA binding proteins, Weliky Conaway said, noting that here the semi-quantitative nature of the mass spec workflow was key.
"The combination of the MudPIT mass spec and their semi-quantitative [analysis] worked really nicely to allow [us] to pull out the population of proteins that were enriched specifically on DNA fragments that contain promoters of interest or under conditions where – as we did – you spiked in a transcription factor that you know is an important regulator of the promoter you're interested in," she said.
As the researchers move on to larger and more complex DNA regulatory systems, this challenge will likely increase. However, Conaway noted, the high throughput enabled by the mass spec-based approach works to mitigate this issue.
In terms of non-specific background binding "one of the key things MudPIT [analysis] allows" is high throughput, he said. "The key thing for eliminating background is having a large number of runs in the database. We generated probably 500 to 1,000 [mass spec] runs. So we have a reasonable idea now about what proteins are always background, what proteins aren't background."
Being able to confidently make such determinations is key to keeping researchers from spending resources chasing down bad leads, he noted, "because actually the hard work comes once you've decided something might be [a] real [interactor] and you want to confirm and explore it."
Weliky Conaway added that increasing mass spec sensitivity should also help researchers better eliminate background binding, as well as identify more low abundance and transient interactors.
"I think [increased mass spec sensitivity] is going to become incredibly important for this approach, both for being able to detect specific things in a large background and also in terms of being able to identify regulatory proteins that interact only transiently with the DNA sequence or DNA binding proteins," she said.