A team led by scientists at the Ecole Polytechnique Fédérale de Lausanne has developed a selective-reaction monitoring mass spec workflow for absolute quantitation of transcription factors.
With the method, detailed in a paper published this week in Nature Methods, the researchers tracked levels of multiple transcription factors including PPARγ and RXRα in terminal differentiation of mouse 3T3-L1L1 pre-adipocytes and used this data to create a model of DNA binding for PPARγ.
As the authors noted, despite the importance of transcription factors to gene regulation, there is relatively little quantitative data on the proteins, in large part due to their low abundance and the resultant difficulties inherent in measuring them.
Indeed, said Marc Moniatte, head of the EPFL's proteomics core and author on the study, when the researchers first began looking for assays to their proteins of interest, they found so little information that they decided to start the process from scratch.
"When we started [several years ago] we really couldn't find any good information about the proteins we were looking at," he told ProteoMonitor. "For all the sequences we tested, there were only [assays to] about three or four peptides available at that time."
In addition to the difficulties presented by transcription factors' low abundance, the researchers also found many of them behave poorly during mass spec sample prep and analysis, Moniatte said.
"The abundance definitely plays a role, but now that we have [assays for] about 200 of them, we have the feeling that it's also the sequence of these peptides that makes them difficult to detect," he said.
For instance, he said, for many their stability in solution over time appeared poor. Others, he noted, flew poorly in the mass spectrometer, making their detection and quantification a challenge.
In fact, Moniatte said, even when using a high-end instrument like Thermo Fisher Scientific's Orbitrap Elite for their discovery work, the researchers were largely unable to detect transcription factor-specific tryptic peptides.
This, along with the general lack of existing mass spec assays for transcription factors, led the researchers to use an in vitro protein expression method to generate full-length versions of their target proteins that they could then use to identify proteotypic peptides for SRM-MS analysis.
They also used this in vitro expression system to create full-length internal protein standards to enable quantification of the transcription factors of interest. This approach, Moniatte noted, allowed the researchers to generate standards shortly before their intended to use, offering a way around the aforementioned stability issues.
"The idea was to express [the standards] a few days or even the day before [the mass spec analysis] to have them at hand without having to store them too long," he said. "That can be quite beneficial to the robustness of the assay."
Implementing the assays on a Thermo Scientific TSQ Vantage triple quadrupole, the EPFL researchers have managed to quantify transcription factors at levels as low as 250 copies per cell, although, as Moniatte noted, this is for "a well-behaved protein."
"It is still quite protein dependent, and some are very difficult to get," he said.
In the Nature Methods study, Moniatte and his colleagues managed to simultaneously quantify up to 10 proteins, which, he said, is roughly the limit of the assay's multiplex capabilities given the low abundance of their targets.
"It's not necessarily a good idea to multiplex too much because then you are introducing too many other peptides that can impair the sensitivity," he said. "So in this case we think that 10 is a good number. It's not very high-throughput, but it is a way to get the information in a precise way."
The researchers are in the process of validating assays with increased multiplexing capabilities.
With the assays, the EPFL team investigated the relationship between genome-wide DNA-binding events as measured by previously published ChIP-seq experiments and the quantity of the relevant transcription factors present, looking at the expression of the proteins PPARγ and RXRα over time during terminal differentiation of mouse 3T3-L1L1 pre-adipocytes.
Using the PPARγ expression data, the researchers developed a model for predicting this protein's DNA-binding events, using the information from their quantitative SRM assays to improve on similar previous thermodynamic models of protein-DNA interactions that relied on inferred transcription factor expression data.
The results they obtained "were quite close to what has already been published in the literature," Moniatte said. "So we think that [the model] is quite good."
Moving forward the researchers aim to apply the assays to a variety of additional projects, he said. "Now that we have these assays and know how to develop them so they are good enough in terms of precision, we have many questions we would like to test."
Currently, he said, they are using the assays to study the effect of perturbing pathways linked to various transcription factors "and looking at the feedback loop." They are also investigating the role of transcription factors in circadian rhythm, studying the variations in copy numbers of certain factors over the course of this cycle.