Phosphorylation is an important modification of proteins because in many cases it controls a protein’s function or activity. It is also of interest to know which proteins are phosphorylated at a particular time. ETH Zurich researchers have now been able to show that none of the known methods for isolating peptides, i.e. fragments of proteins, is sufficient on its own to detect a phosphoproteome of this kind. The paper was published in the scientific journal “Nature Methods” in the week beginning 12 February 2007. Scientists in the same research group were recently able to describe a new method to isolate and identify phosphorylated peptides. It supplements previously known methods.

Most proteins are more than a three-dimensional folding chain of amino-acids. Thus they often undergo further grafting after their initial production from DNA via messenger RNA to form an amino-acid chain. An important grafting of this type is phosphorylation, for which the slang ‘post-translational modification’ is often used,. Phosphorylation involves the attachment of phosphate groups to particular amino-acids. This modification has an effect on important biological processes such as cell growth, cell division or signal transmission.

Because these processes are very complex, there is particular interest in knowing the phosphorylation state of not just one protein but of as many as possible, and certainly in knowing the system state of the grafting process. Carrying out a phosphoproteome analysis of this kind looks relatively easy on paper: proteins from a cell must first be collected and then split into peptides by enzymatic digestion. This cleaving is because peptides can be identified and quantified more easily and in larger amounts in a mass spectrometer than entire proteins. However, the phosphorylated peptides must be isolated before the mass spectrometer and its downstream analysis software enables the peptides to be identified. That means there is need for a method that selectively separates phosphorylated peptides from the non-phosphorylated ones without making any distinctions as to which proteins the phosphate groups are attached.

Different methods, different results

However, this is the very point at which practice and theory part company, because no such method exists. Doctoral students Bernd Bodenmiller and Lukas Mueller together with their colleagues in the research group led by ETH Zurich Professor Ruedi Aebersold of the Institute of Molecular Systems Biology (1) have now systematically demonstrated such a method, eliminating this obstacle. Their paper was published on 11 February 2007 in the on-line version of the scientific journal “Nature Methods”(2).

For their study the researchers compared three different methods frequently used for phosphopeptide isolation: one chemical method and two methods based on affinity for metal ions. Proteins from the model organism Drosophila melanogaster that had been digested into peptides using the enzyme trypsin were analysed. The results showed that it was possible to isolate the same phosphopeptides with high reliability by using the same method. However, an intercomparison of the methods revealed that each covered a different segment of the phosphoproteome, and consequently the individual segments overlapped only partly. Thus in a mass spectrometric analysis of 19,000 recorded signals that were considered to be relevant, only 500 occurred on the basis of all the isolation methods.

If the phosphopeptide isolation method differs, this is visible in the mass spectrometer’s chromatogram. The differences between the lines are the result of the different isolation methods that were applied to the same initial material. (Photo: Institute of Molecular Systems Biology) large

The scientists conclude from this finding that a single isolation method is inadequate to carry out a comparative analysis of the phosphoproteome. In a media report by the Nature Publishing Group, the author thinks this warning should help the research community not to overestimate the predictive ability of a particular method.

Solubility as a solution

But even the individual methods themselves are not yet up to full strength. Quantification in particular needs a lot of work still because phosphorylated peptides often occur at low concentrations. However, Aebersold’s group recently published another paper in “Nature Methods” (3)showing improvement in the chemical bonding method.

The scientists succeeded by using soluble polymers for the first time to bond to and increase the concentration of the phosphopeptides in a proteome analysis. Bonding to the polymers and subsequent detachment was combined with labelling the peptides. The new method works regardless of the particular amino-acid of a protein to which the phosphate groups are attached. Thus phosphorylations on all the three corresponding amino-acids tyrosine, serine and threonine are recognised.

Using several methods in parallel

To obtain even more specific information, the researchers enhanced the method by adding another step. By using antibodies against tyrosine, peptides with this phosphorylation site were enriched. This enabled 75 tyrosine phosphorylation sites and 80 serine/threonine phosphorylation sites to be identified in human leucocytes (T-cells). They also succeeded in following qualitatively and quantitatively the changes in the recorded phosphoproteome caused by a specific treatment of the cells.

One overall fact is clear to Ruedi Aebersold: the individual methods can be optimised in themselves but still remain complementary. That is why several methods must be used in parallel to achieve the most complete possible analysis of the phosphoproteome.