COLLEGE STATION — Profiling, interrogation and surveillance are great tools — especially when the detectives are genetics researchers seeking an understanding of essential cellular processes.
An international team of scientists have developed a method of measuring complex genetic interactions between mutants in a protein complex, gathering thousands of data points to create a “fingerprint.” That, in turn, enabled the team to profile and predict behavior at the cellular level. The study is in this month’s edition of the journal Cell.
“Our long-term goal has been to look at how genes are turned on and off, and then possibly how to regulate that,” said Dr. Craig Kaplan, assistant professor of biochemistry at Texas A&M University and Texas A&M AgriLife Research in College Station. “It turns out that with the genetic profiling, we were able to predict the action of several mutants without having to measure every one, because we know the relationships between them.”
Kaplan collaborated with Drs. Nevan J. Krogan and Christine Guthrie, both with the University of California-San Francisco, as well as with researchers from the University of British Columbia-Vancouver and the University Medical Centre Utrecht-The Netherlands.
The study involved Saccharomyces cerevisiae, popularly known as baker’s yeast, commonly used in research. However, scientists believe the results are predictive of what would be seen if studying the same process in human cells.
“We had structural information (about the yeast), but it’s a very large enzyme, so we wanted to know what the contributions were from its different parts,” Kaplan said.
Hundreds of proteins are involved in gene expression for any particular gene, he explained, and they all converge on one enzyme called RNA polymerase II.
“I took a genetic approach to identify mutants in different parts of this enzyme that maybe had different effects in cells,” Kaplan said. “I wanted to look at different mutants to understand possibly how this process was regulated and what types of things go wrong when you have different mutations.”
Kaplan generated a large number of mutants all over the protein complex, but didn’t have a sophisticated way to understand which mutants were similar.
“Maybe they were similar because they did the same thing even though they were in different parts of the protein, but we didn’t know which mutants were different,” he said.
Kaplan teamed up with Krogan whose lab developed the point mutant E-MAP, or “pE-MAP,” approach which allowed them to get thousands of data points for every mutant.
“Instead of crossing a single point mutant with one or two different mutants, pE-MAP lets us cross each one with more than 1,000,” said Krogan, who also directs the California Institute for Quantitative Biosciences at the University of California-San Francisco. “This gives us 1,000 data points for each mutant, which we then use to build our high-resolution profiles.”
The team used this approach on 53 RNA polymerase II mutants crossed with more than 1,100 other mutants.
“We had at least that many data points, and those are what make the profile,” he said, comparing it to the pixels of digital photography — more pixels means a higher resolution, clearer picture to see,” Kaplan added. “Those 1,100 pieces of data for one mutant are like a fingerprint that says this is specifically how one mutant looks when we test its genetic interactions with 1,100 other mutants. Then we have a profile that has many, many, many more measurements than just a couple.”
Hannes Braberg, Krogran’s graduate student who performed much of the work in the lab, explained why the process was necessary.
“Until now, the only way to get similar information was to deactivate or ‘knock out’ specific genes within an enzyme and observe the impact,” he said.
“But RNA polymerase II is so critical that deactivating even one gene often kills the cell. So instead of knocking out the genes, we mutated them.
“It’s just like you have a scratch on your arm, but you didn’t lose your whole arm,” Kaplan said. “With this method, we didn’t lose something that was essential, we just made a very specific defect.”
The team then cross-examined all of the mutants to find which ones had similar profiles with the understanding that the location of the mutations on the protein didn’t matter as long as the team could link two mutations with similar behavior in the entire complex.
Once the profiles were created, the team compared the mutants’ “fingerprints” to determine which were similar and which were distinct. The combined genetic information helped determine how well cells grow when they have the two similar mutants versus how well do other cells grow when they had two different mutants.
“Does the cell grow very well or is the cell sick?” he said.
The team found that some of the mutants had a type of defect that they otherwise would not have thought to look for — segregating chromosomes to daughter cells.
“RNA polymerase transcribes every single gene in the cell so it could be connected to any process, but we found that some mutants in RNA polymerase can have a specific defect in this process,” Kaplan noted. “We would not have necessarily known to look for that, and yet we were able to see a connection to chromosome segregation out of our assay.”
The researchers also compared the biochemical information for some RNA polymerase mutant enzymes to the profiles to see how the mutants affected the basic properties of this enzyme, such as transcription related to the genetic interactions.
RNA polymerase transcribes DNA to make RNA by turning the gene on and making a copy of it, building together blocks of RNA in a strand based on what the DNA tells it. This is done at a certain rate, Kaplan said.
“When we started to look at some of our mutants biochemically, we found that some of our mutants were slow, meaning when we had a wild-type enzyme it transcribed at a certain rate, but we also had mutants slower in this process that are biochemically defective,” he said. “We had also found that another class of our mutants was biochemically fast. This class had a gain of function, because it was doing something better than it normally does.
“And when we looked back at our genetic profiles, we found that the slow mutants group tended to have the same genetic profile — they had the same types of interaction, were behaving similarly and had the same types of measurements. We started to see a pattern emerge.”
The fast mutants, however, behaved differently, Kaplan added.
“They had different interactions with a different set of mutant genes and that told us that maybe some of the genetic interaction we were observing was related to the biochemical defects that we were discovering,” he said.
“So the fact that we have biochemical data on a few mutants, but we have genetic profiles for many, many more shows that we can now predict which mutants should be slow and which should be fast. We can actually make predictions about all the rest of the mutants that have a similar profile, and we don’t have to measure every one, because we now know that they have a relationship. And that’s very, very powerful.”
The team discovered yet another phenomena in the study involving the rate of transcription.
At the same time the transcription is taking place, another process – splicing – occurs, Kaplan added.
As many mRNA are transcribed, splicing is necessary to “cut out” pieces of the mRNA that can interrupt or alter what protein is encoded and stitch back together what remains.
“There’s evidence that if you alter the rate by which the RNA is made by altering the rate of transcription then the splicing process is affected because it happens at the same time,” he said. “If it goes fast, maybe the process becomes less efficient. If the transcription process is making an mRNA more slowly, the splicing machinery has more time to recognize the appropriate elements, so maybe the splicing reaction will be more efficient.”
Kaplan said scientists had never been able to look at the splicing process with the fast polymerase mutants, nor had they been able to use multiple slow polymerase mutants that have specific, well-defined biochemical defects.
“When Christine’s lab looked at the splicing process in vivo, we found that when you slow down transcription, the splicing process gets better and is more efficient,” he said. “But we also found that when transcription is sped up with our fast polymerase mutants, the splicing process becomes more inefficient. This response to speeding up transcription was predicted by previous proposals (in research) but had never been observed before.”
“This whole approach would be valuable to study any complex or any sort of complicated molecular machine where you don’t necessarily have to pick the exact right mutant to study,” Kaplan said. “You can make or generate hundreds of mutants in a particular complex that may be very complicated and it may have lots of functions, but because you can generate these high resolution profiles — even if your protein touches every single gene in the cell, which is what our protein does — we’re still able to get information out of it and we are still able to de-convolute that information and generate something meaningful. It’s a powerful system-level approach to understand the function of multi-functional complexes.
“This is a process that if we understand how it works at the molecular level, we feel like it will give us a much better handle to predict how things go wrong when they do go wrong or what is the mechanism of how things go right during development or to an organism properly responding to its environment.”