Looking at the big picture

Professor David James and his team are assembling a comprehensive atlas of the phosphoproteome of insulin that is revealing new insights about diabetes.

Insulin is a remarkable molecule, its graceful structural symmetry hinting at the complex role it plays in maintaining metabolic equilibrium. Its function is intimately linked to the process of phosphorylation, which sparks of a cascade of activity within our cells, ultimately enabling glucose to enter the cell from the blood stream.

Understanding the complexities of this phosphorylation process has been the focus of Professor David James’s career, not only looking at how it works, but also how it can break down, causing diseases such as type 2 diabetes. The intricacies of protein phosphorylation and kinase signalling are convoluted, to say the least, and most researchers have tended to hone in on a small chunk of the process in order to bring it into focus. However, it can be difficult to get a handle on the full picture, which is something that James is hoping to correct.

James and his team at the Garvan Institute’s Diabetes & Obesity Research Program have been labouring to develop a comprehensive atlas of the insulin-responsive phosphoproteome, which covers all the phosphate-containing proteins in adipocytes, or fat cells. It is hoped that this atlas will provide the big picture context that will help make the details of insulin’s function become clear.

The three-year effort to compile the atlas was driven largely by a PhD student in the group, Sean Humphrey, who took advantage of the ongoing advances in proteomics, mass spectrometry and bioinformatics to identify and characterise as many phosphoprotein changes as he could find in adipocyte cells following insulin treatment.

This relatively new approach of phosphoproteomics is providing novel information on the phosphorylation-based signalling pathways activated in a given cell under given conditions, which not only tells researchers more about the biological system being studied, but could also lead to new or better drug targets.

According to James, a major strength of their study lay in its highly quantitative nature. The impressive extent of insulin-regulated changes in different phosphoproteins measured in their cells was enabled largely by the sensitivity of today’s mass spectrometers and techniques for phosphoprotein labelling.

“We seem to have a fairly complete and comprehensive dataset,” says James. “It comprises more than 37,000 phosphorylation sites, which I believe represents the largest single-cell phosphoproteome reported thus far.

“And, amazingly, in response to insulin, 15% of those phosphorylation events are changed by more than two-fold. It is a humbling dataset that really demonstrates and confirms how insulin works. We see all the expected things plus some really exciting stuff.”

For example, the team has identified a completely novel substrate of Akt, which is an important kinase in multiple cellular processes including glucose metabolism, with topological features that suggest how Akt itself could be regulated. Indeed, James will highlight this particular finding when he speaks at Lorne in February.

“The dataset also turned up a whole lot of other events of as yet unknown significance. Finding out exactly what it all means will certainly keep us busy for a long time yet.”

Tracking insulin’s timetable

The team also looked at the temporal nature of insulin action by analysing multiple time points after insulin addition, from 15 sec up to an hour, and this approach turned out to be incredibly revealing. In fact, James was quite surprised by the heterogeneous nature of the changes in phosphorylation events across the time scale.

“Some things appear almost as fast as you can kill the cell,” he says. “By 15 seconds they are on, and these tend to be the main signalling players that we already knew about, such as Akt. However, surprisingly, a lot of the other things destined to come on are definitely off in the early stages, and then boom, suddenly they turn completely on. So, it is sort of like a domino effect, and some of the time gaps from one thing turning on to the next were longer than expected in signalling terms.”

The temporal aspect of James’s approach is important in terms of distinguishing various enzyme-mediated events such as phosphorylation (kinases) and dephosphorylation (phosphorylases). Indeed, one of the biggest issues in signal transduction has always been trying to match kinases with their substrates for a given function or endpoint, because those belonging to the same family often have overlapping consensus motifs.

“Temporal profiling therefore really adds another dimension to deciphering these insulin-related phosphorylation events and enables us to resolve some of these relationships that we couldn’t before.

“This time dimension of biological processes has become one of the stable horses of systems biology. You can’t just look at something and compare it to something else at a particular point in time or place. You have to watch things evolve over time to start to unpick the puzzle.”

James admits that this phosphoproteomic approach has certainly broadened his thinking into the realm of systems biology, a term commonly used to describe the study of whole body processes.

“This ‘big data’ is so impressive and exciting because it truly is a completely new way of doing science that requires a rearrangement in one’s way of doing and thinking about research. Even aspects like just being able to handle the damn stuff - that volume of data all at once - looking at it, analysing it and working out how to present it to everyone else in a way that makes sense.”

At the moment, James and his team are trying to take a step back from the enormity of the data and avoid getting caught up in the power of the technology. They instead want to focus on those findings that fit in best with the group’s ongoing research activities and questions, and explore them in more detail.

Getting personal

According to James, there is even plenty of data to go around, and James expects that once the work is published there will be many other researchers interested in applying some to their own activities and aims or entering into collaborations with James to work on molecules of interest to both groups.

“Actually, Sean has already set up a really nice collaboration with Jean Yang’s group at Sydney University. She is a biostatistician with expertise in computational biology. Together with a guy from her group, Sean worked on how to analyse and visualise the data, even things as seemingly simple as how to show the data to others.”

This is one of the important issues in systems biology, says James. “You have all this information and then have to work out how to show it to someone and make it stand up and say the things you think it says. It is not so easy.”

One of the things James is really excited about is taking these new high-tech results almost directly into the clinic. “I believe that we are at somewhat of a turning point in the history of medical care and health delivery that will change some of the conventional procedures of symptom-based disease diagnosis. Indeed, it is already happening in some areas.

“More and more, doctors will have all of this new information at their fingertips: a patient’s genomics profile and potentially the relevant phosphorylation status, signal activation data etc. And I can see the possibility of taking data such as ours straight into the clinic to provide clinicians with a much greater capacity to make more specific and sensitive diagnoses.

“The real question all of this brings up in the clinical arena is: what actually is disease? Take diabetes, for example, which for the most part we think of as a single disease entity, when in fact it may be a thousand different diseases whose course depends on many things, such as the environment that a particular person is exposed to and so on.

“Where I see these big dataset types of experiments being so powerful when applied in the clinic is in helping us work out this personalised aspect to medicine, especially in complex diseases like diabetes.

“And genomics alone, in my opinion, will not be the one answer. You can’t just look at gene type and make predictions. You have to look at multiple parameters that define the biology of a system. This will be the real power of true systems biology. As a case in point, here am I, having spent the last 20 years of my research life poking around with a few molecules and then we do this relatively short project and end up with the data we have, and you just think ‘Oh my God, so that is how insulin really works!’”