Plants, proteins and proteomics

Profiling proteins is a challenging undertaking. But new research is providing new insights into protein structure and function, reports Susan Williamson.

Harvey Millar believes the way to create hardier plants is through proteomics -- and apparently the federal science minister agrees.

Millar, a professor of plant biochemistry at the University of Western Australia, won the 2005 Science Minister's Prize for Life Scientist of the Year for his work, which uses proteomics to reveal how plants function.

His scientific career began well before the language and technology of proteomics was born. But Millar has since embraced the concept of the proteome and extended it to the protein composition of his pet organelle, the mitochondria, to the rest of the cell and the whole plant.

After doing a postdoc at the University of Oxford and prior to that a PhD at the Australian National University in Canberra, Millar moved to Perth in 1999. He is now embedded in the recently formed ARC Centre of Excellence for Plant Energy Biology at the University of Western Australia (www.plantenergy.uwa.edu.au).

He and his colleagues have contributed to better understanding of how plants work, by looking at plant genes and gene transcripts. Now, they are employing proteomics to look at plant proteins and reveal where the gene products are located within a plant and, more specifically, within a cell.

"There's about 30 years worth of biochemistry that's been done to purify different sub-fractions of plant cells, of which the mitochondria is one fraction," Millar says. "What we can do now, using proteomics, is to analyse those cell fractions so we have a good idea of what gene products are located in which particular locations in the cell."

Plant-specific proteins

Millar's team is interested in primary metabolism in plants. They are particularly interested in three main organelles -- mitochondria, chloroplasts and peroxisomes. The generation of metabolic energy is a major activity of all cells, and these three organelles coordinate a fairly contiguous set of reactions in respiration, photosynthesis and nitrogen assimilation.

Mitochondria are involved in the breakdown of carbohydrates to generate energy in the form of ATP, whereas chloroplasts capture energy from sunlight and generate ATP as well as the synthesis of carbohydrates. Peroxisomes contain enzymes involved in different metabolic pathways, including the breakdown of fatty acids and the metabolism of by-products of photosynthesis.

In addition, mitochondria and chloroplasts contain their own genomes. The mitochondrial genome in plants is much bigger than in animals, encoding about 50 proteins, although most of the protein products found in mitochondria are encoded in the nucleus and imported into the mitochondria via N-terminal targeting sequences. The proteins are post-translationally targeted to the mitochondria by virtue of these targeting signals, which are essential for their entry into mitochondria.

"What we are doing at the moment is try to answer the question 'what are the nuclear gene products that go to mitochondria?'," says Millar. "There have been various attempts to determine what that protein set is in the nuclear genome, using bioinformatic approaches to look in genes for what we believe to be a targeting sequence and building up sets of proteins in that way.

"The problem is we don't know a lot about the 'language' that is spoken in the cell to send proteins to particular locations. We have some idea but depending on which targeting prediction program you use, you get different proteins predicted to go to mitochondria, which doesn't help a great deal when you're trying to work out a list."

This is where proteomics can help. The mitochondrial sub-fraction of a cell, for example, can be taken and refined to a high degree of purity enabling the proteins in the sample to be identified.

"We've been doing that and we currently have a set of about 500 proteins," says Millar. "We can then take that set and, using bioinformatics approaches, compare it to similar sets that have been identified from yeast and human and mouse mitochondrial extracts and start to determine which mitochondrial-located proteins are conserved at a gene level across those organisms, which ones aren't, and which ones are unique to each organism."

When Millar and his team did this comparison they found that the highly conserved proteins were those involved in processes mitochondria have long been known to conduct, such as electron transport chain proteins, and proteins involved in the TCA cycle -- as Millar says, "basically, the primary machinery of making ATP is conserved".

The researchers also found that mitochondria from different organisms contain many other proteins that are either unique to that organism, or the abundance of them is very skewed in one organism compared to another.

Millar's team is looking at some of these pathways that are unique to plants. Identifying these plant specific proteins will give insights into how the electron transport chain is regulated and what it is capable of achieving in plants compared with animals.

"For example, plants make a number of vitamins in the mitochondria, such as vitamin C, some B complex vitamins like folate and biotin, which are not made in mitochondria in humans," says Millar. "There is a whole array of other carbon metabolism-related proteins plants have that animals don't, because animals don't carry out processes like photosynthesis."

Protein complexes involved in the respiratory electron transport chain in mitochondria have been heavily researched in mammals, but not very much work has been done on their counterparts in plants. "This is partly because people thought they would just be the same," says Millar. "But when we have pulled them out and looked at them, we find some are very highly conserved, but others are very plant specific."

Function with location

One of the problems with this work is purifying the mitochondria well enough to be sure there aren't any contaminants from other fractions in the cell in the sample.

As well as trying to purify mitochondria more completely, Millar's team is looking to identify the locations of all proteins in the plant cell. Knowing the physical location of all proteins makes it easier to identify protein contaminants in a mitochondrial fraction, since a given protein might be found to be abundant somewhere else in the cell.

"We have been bringing together lots of sub-cellular proteomics studies from around the world so we can build databases for proteomics that show the locations of proteins within plant cells right across all the different organelles," says Millar. "You can then get a whole cellular picture of protein locations, which we think is a pretty fundamental piece of information that's surprisingly absent for plants."

To confirm the location of proteins defined by proteomics, research using green fluorescent protein chimerics with a protein of interest enables live cell imaging to be conducted, and when the protein is expressed it can be tracked and its location in the cell observed.

"If we can work out where proteins go, even for the ones with an unknown function, it gives us a cellular environment in which they are operating and we can start to look at things like co-expressed groups of proteins that go to the same place," Millar says. "Then you start to get down to things that look much more like biochemical pathways in some way, so it breaks up the problem of trying to work out the function for proteins with unknown function by looking at location."

Millar and his team have incorporated this sort of data with protein location by mass spectrometry. In addition, they have incorporated about eight different bioinformatics targeting-prediction programs that attempt to decipher where proteins are in the cell. By incorporating the bioinformatics, they can compare prediction to experimentation of where a protein might be located.

"That's helping us," says Millar. "It's helping other people more broadly in that they can see all the data, and it's helping us to start to assess some of the proteins we have found in mitochondrial samples and ask the question 'are they really mitochondrial or are they contaminants?'. It's a slow process of going through all the data, and that's where we are at the moment."

Pick of the crop

Millar says that as well as expanding knowledge in plant science, his team's research in energy conversion and generation in plants has applications to agriculture.

"We have a variety of plant varieties in the Centre [of Excellence for Plant Energy Biology] that have altered capabilities as whole organisms because we have modified organelle functions, and that ranges from mitochondrial functions to chloroplast functions to peroxisome functions.

"We are trying to understand how a cell regulates its primary metabolism. If we can understand that we're in a position to alter a plant's decisions, if you like, at the level of energy generation."

By altering a plant at the level of the cellular organelles -- what Millar calls "tinkering in the engine room of how plants work" -- a plant's robustness can be improved giving it more energy to cope with a specific situation.

Some of the plant lines his team has generated in the model organism Arabidopsis are more drought resistant, flower at different times, grow faster, or cope better with salinity. A lot of this work involves researching a plant's responses or defence mechanisms to oxidative stress. Millar's team has worked out which proteins are most damaged by oxidative stress and what transcriptional and translational responses organelles make to oxidative stress.

"A couple of years ago we found that mitochondria use the same major defence mechanisms employed by photosynthesis against oxidative stress to avoid the production of free radicals. They both use vitamin C as an anti-oxidant, like humans do," says Millar. "In fact it's even the same genes, so these genes produce proteins that are dual-targeted to the chloroplasts and the mitochondria."

These organelles are the sites of production of reactive oxygen species in oxidative stress because they contain the electron transport chains, oxygen is a substrate or a product of the electron transport chains, and this chemistry tends to lead to free radical production when things go wrong.

Vitamin C is therefore one way in which plants coordinate their response in these organelles.

"Because all the Arabidopsis genome has been sequenced there is a very high success rate of identifying peptides with proteomics in this plant -- we have all the data to pattern-match against the genome," Millar says. "It's not a crop species at all, so the challenge -- to apply this to agriculture, to take this foundational knowledge in a model plant and translate that into crop species of interest and see if you get the same phenotypes -- is a major undertaking down the line."

Millar sees the work of the UWA centre as sitting behind major research institutions like the CSIRO and Research & Development Councils, who are much better equipped to take this work into crop species in the field. "We are essentially defining targets from our perspective in the biology for these types of organisations to take forward," he says.

Detecting low abundant proteins

Millar's group has obtained Large Infrastructure Grants from the ARC over a number of years now and has used them to purchase proteomics infrastructure. The team now has a number of mass spectrometers at its fingertips.

"We mainly use liquid chromatography and tandem mass spectrometry," says Millar. "And we use a combination of gel-based systems and gel-free systems to look at the proteins. You don't find everything on gels, membrane proteins are problematic in gel-based systems, so we have to use non-gel-based systems to look at those because of protein solubility problems."

To analyse low abundant proteins and concentrate very small samples, Millar says his group is using very low flow rates with liquid chromatography. They have nano-flow rate HPLCs entering into mass spectrometers, which Millar believes is the way things are going as it helps in getting high sensitivity in mass spectrometry.

The centre recently purchased a new piece of equipment, called free flow electrophoresis, which enables isoelectric focusing, like the first dimension in a 2D gel, but in liquid, so the researchers can fractionate a sample based on isoelectric point in 96 fractions. This allows them to then look at a single fraction separated from high abundant proteins and therefore revealing the lower abundant ones.

"We suspect that there might be something like 1500 proteins that can be targeted to mitochondria under particular circumstances," Millar says. "Many of them are probably extremely low in abundance and they are going to be the hardest to find but possibly the most interesting."

The most abundant protein in plants is Rubisco, which does the CO2 fixation in chloroplasts -- comparable to albumin in blood samples, Rubisco can make up to 20 per cent of cellular protein in whole plant leaves.

Down at the mitochondrial level, Millar says the proteins that dominate things are mainly electron transport chain complexes, such as some of the major carbon catabolising enzymes.

The researchers' aim is to remove these high-abundant proteins so they can look at some of the lower abundant proteins, such as those that regulate the mitochondrial genome, which Millar says are often very low in abundance.

"We are also doing work on the post-translational modification of proteins in the mitochondrial proteome. We are looking at phosphorylation and have found several putative protein kinases and we are quite interested in what their targets are -- we are doing this by mass spectrometry to try to identify the sites of phosphorylation."