Networking at the molecular level rather than just at the level of organisms within the ecology is a big subject in biology. Understanding this interaction is an important goal of systems biology, but the aim is as difficult to attain as it is noble. For example, often all that could be found out in protein interaction studies was whether the individual molecules being investigated together with a target protein reacted or did not when an inhibitor was added. However, one did not know whether these reactions occurred because of a specific reaction with the target protein. Quantitative estimate also fell by the wayside again and again.

However, ETH Zurich researchers from the Institute of Molecular Systems Biology have now developed a new method (1) that enables specific interactions to be recorded, a quantitative analysis to be made and the phosphorylation of proteins to be determined, all in a single step. The paper is published in the March issue of the scientific journal “Nature Biotechnology” (2) .

A mass-spectrometric analysis forms the starting point of their new method. Mathias Gstaiger, a co-author of the paper, explains: “We chose this method because mass spectrometry allows the maximum number of proteins to be recorded simultaneously.” His colleague Lukas Müller adds that the mass accuracy and thus the resolution of this method were greatly improved in the last few years.

However, when a measurement is performed, good analysis software is needed to put the signals into a form that can be interpreted and/or to filter out the relevant signals. This is done using the “Superbrain” software, also developed at ETH Zurich. However, a “Superbrain” is still no guarantee of showing the specificity of the bonds.

The ETH Zurich scientists demonstrated this with a special trick. They took the protein whose binding partners were to be discovered and carried out a dilution series. This caused the intensity of the corresponding signals in the mass-spectrometric analysis to decrease. Oliver Rinner, another author, explains: “This allowed us to define seven specific binding partners for our model protein FoxO3A from a soup of 20,000 proteins. Gstaiger points out that this finding was anything but self-evident, since mass spectrometry was considered only a semi-quantitative method until then.

Working as a team: Ruedi Aebersold, Matthias Gstaiger, Lukas N. Mueller, Markus Müller and Oliver Rinner (anticlockwise from lower left, shown in a diagram) have developed a new analytical method for protein interactions which provides information about peptides that are specific for particular proteins. (Photo: L. Mueller)

However, that is still not the full story of the spectrogram interpretations. Thus the ETH Zurich scientists were also able to recognise the phosphorylation status of the proteins, i.e. to determine whether or not phosphate groups are attached. Markus Müller, a bio-computing scientist and co-author, explains that this can be read out from the systematic shifting of the signal peaks.

The new method creates a Master Map as an overall picture of the experiments, for FoxO3A in the instance of the model case carried out by the ETH Zurich researchers. According to Gstaiger “This map represents a fingerprint.” It also contains data about how a protein responds to changed conditions. The ETH Zurich scientists analysed this for FoxO3A by inhibiting a particular phosphorylation.

A Master Map can also be a basis for further studies, for example, in which an individual reaction is investigated. That’s why the ETH scientists want Master Maps to be freely accessible via the Web and to be freely available on new data bases. One aim at the Institute of Molecular Systems Biology is to prepare further Master Maps of this kind for entire protein groups as well. A project in which 50 proteins from Drosophila are to be studied simultaneously is already in progress in collaboration with ETH Zurich Professor Ernst Hafen’s group.

Overall the researchers are convinced that their method has a pioneering role in the analysis of protein complexes. Collaboration in the project was also revolutionary in itself. Gstaiger said “We did not carry it out as a classical paper by one or two doctoral students obtaining additional information from other researchers. It was much more of a team effort by various groups.” Thus it was possible to complete the study in just over one year, whereas it would probably have taken three years using the classical approach.