Cracking the secrets of posttranslational modifications
Cells do what proteins tell them to do. But sequencing DNA or running microarrays won’t reveal a protein’s mandate. During and after translation, enzymes, lipids, proteins, and sugars decorate the amino acids of the newly synthesized protein. As a result of these alterations, proteins encoded by the same genetic sequence often diverge in function.
Researchers have begun to get the lay of the land in terms of what types of modifications occur, and, broadly, their effects. For example, enzymes are activated or deactivated when a phosphate group latches onto an amino acid (phosphorylation). Metabolism can change when an acetyl group attaches to lysine in a mitochondrial protein (acetylation). The addition of a methyl group to histone peptides (methylation) regulates gene expression. And when SUMO (small ubiquitin-like modifier) proteins attach to other proteins (SUMOylation), they affect cellular processes ranging from apoptosis to neurodegeneration.
But there’s still scant knowledge of which peptides are modified and where, how modifications arise, and the effects they have on cells. Mass spectrometry analyses can assess simple modifications like phosphorylation ever more efficiently and accurately, but readily picking up the scent of modifiers like the SUMO proteins is technically challenging. Studies of modified-protein function are in their infancy, but new techniques to synthesize these molecules will pave the way by helping researchers tease out their effects, one modification at a time.
The Scientist surveyed researchers about innovative approaches they’ve devised to answer stubborn questions about posttranslational modifications. Here’s what they said.
Markus Aebi, Professor of Mycology, Eidgenössische Technische Hochschle (ETH), Zürich
How to study proteins modified by polysaccharides?
Many glycosylated proteins—proteins with attached sugar chains—defy study because carbohydrate modifications are wildly variable and difficult to synthesize. Aebi says, “We can sequence large chunks of DNA in a day, but to determine the structure of one carbohydrate we need weeks, if not months.” Yet these modifications are common, as well as vital for health, because mammalian cells use glycoproteins to detect pathogens. To model different types of glycosylation, Aebi found a way to mimic mammalian glycosylation with help from bacteria and a synthetic sugar (Nature Chemical Biology, 6:264-66, 2010).
Aebi and his colleagues cloned an enzymatic pathway involved in glycosylation from the bacterium Campylobacter jejuni, and transferred the sequence into E. coli. Using Campylobacter’s pathway, E. coli generated a variety of novel transgenic glycoproteins. The team then modified the glycoproteins in vitro by removing a terminal segment containing a bacterial sugar and attaching a sugar they had synthesized, a mimic of an N-type glycan common in mammals. “This is glycoengineering,” says Aebi, who has patented the E. coli portion of the protocol.
Although certain glycoproteins, like antibodies and those characterizing ABO blood types, are available commercially, most must be made from scratch. Aebi broadened the palette of glycoproteins for study by allowing researchers to add their own chemically synthesized sugars. The only tricky part now is synthesizing those sugars, says Aebi, adding that the approach “opens up totally new possibilities” by providing a means “to produce glycoproteins carrying natural, full-sized eukaryotic glycans.”
With Aebi’s method, researchers may synthesize molecules for therapeutic as well as basic-research purposes, such as learning how some mammalian glycoproteins protect their host against specific infectious agents. Testing the functionality of engineered glycoproteins isn’t straightforward, however, and Aebi’s lab is in the process of developing a method to ensure that the engineered glycoproteins fold as they should.
Till Bartke, Postdoc in molecular biology, Gurdon Institute, Cambridge, UK
How do nuclear proteins react to modifications in histone peptides?
DNA and the histone proteins it winds around form nucleosomes. Within the nucleus, bundles of nucleosomes relay information to the cell through proteins that recognize bits of the DNA and histones. Bartke wondered how posttranslational modifications of histone peptides altered the composition of such interacting proteins to generate “signals with biological meaning,” he says.
Bartke’s approach was to make his own nucleosomes from modified histones and methylated DNA, and then observe which proteins stick (Cell, 143:470-484, 2010). To make a nucleosome, he and his colleagues first expressed human histone proteins and DNA separately in bacteria, then purified the histone and DNA products. Next, they methylated the DNA and linked methylated peptides to the histones. They then assembled the methylated histones and DNA into a nucleosome particle. Finally, they attached the prepared nucleosome to an agarose bead, soaked the complex in nuclear proteins extracted from human cells, and identified which proteins stuck.
Bartke’s team homed in on one histone modification per histone and several DNA modifications within a single nucleosome, but that’s just the tip of the iceberg. They hope to extend the work to different combinations of histone modifications in multiple nucleosomes.
But the technique isn’t for the faint of heart. Making nucleosomes takes time and biochemical expertise. Each step in the process requires a few purification steps of its own. Including the time needed to prepare the materials, such a project takes several months. “It’s necessary to plot out this experiment far ahead of time,” Bartke warns.
Still, it may be well worth it. In the past, researchers have taken a simpler tack, assessing protein reactions to modified histone peptides alone instead of incorporating the modified peptides into whole histone proteins wrapped in DNA. Bartke says that looking at histones within a nucleosome reveals how three-dimensional structure alters the information passed on by interacting proteins—information you won’t glean from studies of lone peptides.
Anna Kashina, Associate Professor of Biochemistry, University of Pennsylvania, Philadelphia
What determines a protein’s posttranslational state?
Cell morphology and cell migration rely on two different versions of actin. Although the nucleotide sequences of RNA encoding both versions are nearly identical, only one version of the protein tends to contain bound arginine (arginylation). Because nucleotide differences have been found to affect the rate of translation, Kashina’s group decided to see if the difference in posttranslational state was related to the speed of translation.
Kashina relied on a thirty-year-old procedure to infer the rate of translation. Her team estimated the rate at which radioactively labeled amino acids became incorporated into peptides in the ribosome during actin translation. The more labeled amino acids incorporated, the faster the translation. They found that speedier translation meant faster protein folding, which resulted in a slightly different protein structure. The folded conformation of the actin that was translated faster was far more resistant to a type of degradation that can happen after arginylation.
Kashina’s team used a commercial kit for estimating the rate of translation, simply adding a radiolabeled amino acid to those that came with the kit.
Meanwhile, a biotech startup called ANIMA Cell Metrology in Bernardsville, New Jersey, cofounded by Kashina’s colleagues at the University of Pennsylvania, is working on a more direct way to measure the rate of translation by simply timing the process.
Patrick Pedrioli, Program leader, Scottish Institute for Cell Signaling, Dundee
How can mass spectrometry detect complex posttranslational modifications?
Researchers can rapidly characterize simple posttranslational modifications like phosphorylation, acetylation, and methylation with the help of mass spectrometry: digesting the proteins and identifying the resulting peptides by the distribution of ions they emit upon fragmentation. But because SUMO and other ubiquitin-like modifiers are themselves proteins, rather than small compounds like phosphate, they break apart and send mixed messages to standard mass spec analytical software. So Pedrioli created software called SUMmOn, which employs a novel analytical strategy to search mass spec readings (Nature Methods, 3:533-39, 2006).
SUMOylated peptides (peptides with small ubiquitin-like molecules attached) tend to emit multiple mass-spec signatures as they fragment, unlike the single signature emitted by a phosphate group. Pedrioli programmed SUMmOn to compare spectra from a user’s experiment with results from a protein database of various spectra given off by SUMOylated peptides, and identify any matches. In doing so, it provides a score indicating the likelihood that each spectrum contains SUMO fragments and the likelihood that a peptide was SUMOylated. Pedrioli says, “SUMmOn understands the fragmentation of SUMO.”
How SUMOylated proteins change throughout the cell cycle in response to hormones, nutrients, stress or disease are all questions this software could help answer. The method used in SUMmOn is theoretically applicable to any complicated modification, as long as it generates a unique fragment ion pattern, Pedrioli says. And, to improve the efficiency of the hunt, Pedrioli and others are developing techniques to filter out everything but SUMO and other ubiquitin-like modifiers from protein samples so that mass specs are more likely to catch these types of proteins.
SUMmOn is freely available at http://summon.sourceforge.net, and interfaces with most mass specs. However, installing the software requires computer expertise. Users need access to UNIX operating systems and some basic knowledge of UNIX system administration. Pedrioli also says users should perform an independent form of validation, like Western blots, prior to publication, to experimentally ensure that SUMOylated peptides identified by SUMmOn are in fact SUMOylated.
There are still aspects of SUMOylation that Pedrioli’s software (and other techniques) can’t handle well. For example, SUMO and other ubiquitin-like modifiers often form chains when they attach to proteins, but because proteins are digested into peptides prior to mass-spec analysis, these applications can’t detect what peptide the chains are conjugated to.
Correction: Anna Kashina was incorrectly described as an Assistant Professor of Cell and Developmental Biology. She is an Associate Professor of Biochemistry. The Scientist regrets this error, which has been corrected in the text of the article.