Ask anyone in civil service – taking a census isn't easy. Gathering information about a population in a specific region can be a logistical nightmare; would-be respondents are often not at home or simply don’t answer the door.
Sometimes a person ends up in the mix that shouldn’t be; a visitor gets added to the count by mistake, or a clerical error places a person in the wrong residence. In a big, hectic world, accurately tracking where individuals live is a major challenge.
Now imagine carrying out such a project at a cellular level, mapping the human proteome – the entire set of proteins in the human body – amidst the busy molecular landscape of human physiology. Humans contain over 20,000 protein-coding genes, which likely produce over 100,000 distinct proteins. Finding and mapping them all is a massive and often tricky undertaking.
"We don't yet have a home for every single protein encoded by the human genome," said Vamsi Mootha, a senior associate member of the Broad Institute and co-director of its Metabolism Program. “But finding that home is important because one of the first steps to understanding the function of a newly-identified protein is to figure out where it lives.”
Mootha is part of a team, led by Broad associate member Alice Ting, that developed a method that may make this mapping process easier and more precise. The team, which includes researchers from the Broad and Massachusetts Institute of Technology (MIT), found a way to localize large numbers of proteins within sub-compartments of living cells by tagging them with a biological compound that makes them easier to spot.
The group tested the method on the mitochondrial matrix – the winding space enclosed by the inner membrane of mitochondria, the cellular regions often referred to as the "powerhouses" or “batteries” of the cell. The method helped them map 495 proteins to the matrix, 31 of which had not previously been linked to mitochondria. The results were published in the Jan. 31 online edition of Science.
"This method also makes it possible to map the proteomes of other regions of the cell, some of which couldn't be mapped by any other method,” said Ting, who is also an associate professor of chemistry at MIT. “It's a tool for understanding the molecular compositions of cellular regions that were black boxes before."
Ting’s group looked at a series of enzymes – complex biological substances that cause chemical changes in the body – and found that ascorbate peroxidase (APEX), when introduced to a derivative of biotin (a B vitamin that is essential for the activation of many enzymes), successfully tagged targeted proteins in the mitochondrial matrix of live, cultured human kidney cells. To make sure that only the matrix was being targeted, the researchers bound APEX to a compound known to inhabit the region.
The tagged matrix proteins were then identified by mass spectrometry, a tool that determines the ratio between the mass and charge of molecules in a sample, giving away its chemical composition. This work, which was carried out by team members Namrata Udeshi and Steve Carr in the Proteomics Platform at the Broad, used a labeling method called SILAC (stable isotope labeling by amino acids in cell culture) to help distinguish the tagged proteins from other proteins and biological background noise. The team used MitoCarta, a Broad database of previously-identified mitochondrial proteins, to give them confidence that the hits they were getting with their method were accurate.
This new method is more specific and efficient than methods conventionally used to map proteins within the cell. Usually, cells are run through a centrifuge to separate parts of the cell by density, and then researchers pull out the fraction that contains their object of interest. Cells get destroyed in the process, proteins get lost in the mix, and often, like the misplaced souls in the census analogy, proteins that don’t belong in the targeted cellular region find their way into the sample, contaminating the results with “false positives.”
The team said that minimizing these false positives is one of their method’s major advantages.
“With this method, it’s much easier to deal with false positives. With conventional methods, it’s hard to pick out which are bona fide mitochondrial proteins and which are from the background. With this method, we’re able to make a statement about which proteins are contaminants,” team member Namrata Udeshi explained.
Members of the team also pointed to the specificity of the method as another advantage.
“This method allows us to take the general parts list of the proteome and localize the parts to these subdomains within the cell,” Mootha said. “It allows us to figure out not just what state a particular protein resides in, but what zip code it resides in as well.”
Mootha, who is also a professor at Harvard Medical School and Massachusetts General Hospital, emphasizes that, while the proteins themselves will be of interest to researchers who study mitochondria, it is the method that will be of greatest utility to the broader scientific community.
“We’ve made good progress in defining the mitochondrial proteome over the last decade here at the Broad using mass spectrometry and proteomics, so the parts list has been nicely assembled,” Mootha said. “What’s really beautiful here is the opportunity to localize large numbers of proteins in a single experiment.”
The researchers say the method can be generalized. It could be used to explore the proteome of the endoplasmic reticulum, for instance, or other cellular sub-compartments. Ting's team is currently using it to analyze the protein composition of the mitochondrial intermembrane space, which could not be examined at all using conventional methods.
The technology, which can be used to analyze live, cultured cells, may also have future applications that could impact patients.
“Because this method is easy to implement and requires such a small amount of material, it has the potential to be used for proteomic analysis of patient-derived samples,” Ting said. “There could be translational applications in terms of the evaluation of therapeutics and understanding the molecular mechanisms of disease. I think that would be an exciting direction to go in in the future.”