Common Chemistry To Produce Uncommon Results

---

How to Enhance the Detection of Phosphopeptides by Liquid Chromatography -Electrospray Mass Spectrometry

As a subset of proteomics, phosphoproteomics involves detection and characterization of phosphopeptides from biological samples. Because protein phosphorylation is involved in many important cellular processes, such as signal transduction, gene regulation and metabolism, phosphoproteomics has become a very important approach in studies of cellular processes and their regulation.
Liquid chromatography electrospray tandem mass spectrometry (LC-ES-MS/MS) is currently the most common tool in proteomics and it has been employed with great success in many applications.

However, the application of LC-ES-MS/MS in the detection and characterization of phosphopeptides remains challenging, especially for multiply phosphorylated peptides (i.e., those peptides that contain more than one phosphorylated amino acids). However, researchers at the University of Western Ontario in professor Gilles Lajoie's Biological Mass Spectrometry Laboratory (BMSL) have found a potential cause and solution for this problem1.

When biologists and mass spectrometrists explore phosphorylated proteins with modern LC-ES-MS/MS instruments and methodologies, they are often frustrated by their results. More accurately, they are frustrated by the lack of phosphopeptides detected, even when analyzing samples known to contain many phosphorylated peptides. A quick survey of the literature confirms this problem with phosphorylated peptides2. Several theories have been put forth to explain this phenomenon, including the relatively low abundance of phosphopeptides in vivo, phosphorylated peptides' low ionization efficiencies (i.e., protonation) in the positive ion mode electrospray process, loss of hydrophilic phosphorylated peptides in reversed phase LC-ES-MS/MS3 and the unwanted interaction of phosphorylated peptides with the silanol groups of the C18 reversed phase material4.


Recently, the research team in BMSL has discovered another possible mechanism that explains why phosphopeptides are so difficult to detect by using LC-ES-MS/MS: the formation of phosphopeptide-metal ion complexes1. The formation of phosphopeptide-metal ion complexes, which typically contains either Fe(III), Ni(II), or Al(III), dramatically decreases the signal intensity for the protonated phosphopeptides and sometimes completely suppresses it.

Another downfall arising from the interaction between phosphopeptides and metal ions on reverse phase LC columns may also result in the unusual chromatographic behaviour observed for the phosphopeptides, which can make these peptides difficult to be eluted and detected by LC-ES-MS/MS. The proposed mechanism appears to explain well many of the previous results observed in the literature.

When the phosphopeptides were analyzed by ES-MS using direct infusion for sample introduction, the phosphopeptides were detected as protonated ions without any problem. However, when the samples were introduced into the MS by LC, the phosphopeptides were detected mainly as phosphopeptide-Fe (III) complex ions. Since most MS software is programmed to recognize only protonated species for analysis, these metallated ions can be lost in the data handling step. This characteristic alone could explain why phosphopeptides are difficult to detect by LC-ES-MS/MS.

In order to minimize the formation of these phosphopeptide - metal ion complexes, a very classical method was chosen to reduced the presence of metal ions - chelation with ethylenediaminetetraacetic acid (EDTA).

The EDTA was added to the phosphopeptide samples to be analyzed by LC-ES-MS/MS at different concentrations in order to sequester the Fe ions. At a concentration of 25 mM EDTA, the protonated phosphopeptide ions were detected as the most intense peaks for all samples and the intensities of the phosphopeptide-Fe (III) complex peaks were suppressed to below 3 per cent of the protonated peaks (Figure 1). EDTA prevented the formation of phosphopeptide-Fe (III) complex ions and enriched the protonated ions of phosphopeptides.

The benefits of employing EDTA in LC-ES-MS/MS experiments were further evaluated by analyzing more complex samples - the tryptic digestions of two well-characterized phosphoproteins: -casein and -casein. Upon tryptic digestion, these two proteins generate several phosphopeptides that bear one or more phosphoserine residues. It has been reported that the tetraphosphorylated peptide from -casein was very difficult to detect by LC-ES-MS/MS (i.e., it could only be detected at a detection limit 1000 times higher than the monophosphorylated peptide from -casein)4.

When a sample of -casein digest (1 pmol) was analyzed by LC-ES-MS/MS without EDTA, the monophosphorylated peptide was detected easily, but the tetraphosphorylated peptide was not detected at all. Careful analysis of the data revealed that there were some phosphopeptide-Fe (III) complex ions of the monophosphorylated peptide detected, but not for the tetraphosphorylated peptide. This result matched previous findings reported in the literature3, 4.
However, when EDTA was added to the LC-ES-MS/MS system, both the mono- and tetraphosphorylated peptides were clearly detected as protonated ions. The limit of detection observed for the tetraphosphorylated peptide from -casein was 25 fmol - 1000 times more sensitive than previously reported in the literature4. The result from analysis of -casein gave similar results.

Also noticed was a very interesting detail - the chromatographic peak shape obtained for the single-ion chromatogram of the tetraphosphorylated peptide ions was very broad (7 min). From this phenomenon and considering other published results in the phosphopeptide analysis, it was also hypothesized that the tetraphosphorylated peptide was retained on-column by two mechanisms: (1) the well-known reversed-phase mechanism and (2) by complexing with metal ions in the stationary phase material in LC. It has been suggested that the tetraphosphorylated peptides are not efficiently retained on reversed-phase columns because they are too hydrophilic3. However, this conclusion was drawn from indirect evidence (i.e., the phosphopeptide was not detected by LC-ES-MS/MS). In the study, the 'lost' multiply phosphorylated peptides were 'found' after the addition of EDTA.

These results indicate that besides the reversed phase mechanism, phosphopeptides, especially multiply phosphorylated peptides, were also retained on the reversed phase column by a second mechanism based upon a phosphopeptide-metal ion interaction.

From the results, two conclusions could be drawn. First, was the observed formation of complexes between phosphopeptides and various metal ions (Fe(III), Ni(II), and AI(III) when borosilicate tips were used for sample introduction), and that these complexes dilute the signal for the protonated ions. The degree of the phosphopeptide-metal complex formation increases with the degree of the phosphorylation, but is also peptide sequence dependent.

Second, the chromatographic behaviour of the analyzed phosphopeptides demonstrates that these peptides may be interacting with metal ions in the C18 material. The multiply phosphorylated peptides may have such a high affinity for binding to the metal ions in C18 that they do not elute from the LC by a reversed phase gradient alone - the addition of EDTA is required for elution. This hypothesis is supported by the fact that, when large amounts of phosphopeptides are injected into an LC-ES-MS/MS system, protonated ions are easily detected, but they elute as sharp peaks followed by a wide 'hump'. This indicates that two retention mechanisms may be at work, one being the standard reversed-phase mechanism and the other being a phosphopeptide-metal ion interaction. When the amount of injected analyte is large enough, the metal ion sites were saturated and the excess phosphopeptides elute as a sharp peak.

It also explains why the multiply phosphorylated peptides were only detectable above certain amounts during 'standard' LC-ES-MS/MS experiments.After the publication of the findings, it has received much attention and interest from scientists in the field of proteomics. One of the first questions asked is where the metal ions come from. It was not commonly realized that even the purest solvents contains a trace amount of metal ions as contaminants.

In most HPLC grade solvents, individual metal ion content is 0.5 mg/kg. This level of metal ion might be insignificant for certain analytical applications but could greatly affect phosphopeptide analyses by LC-ES-MS/MS. This problem may persist and become more serious as mass spectrometer becomes more and more sensitive.

Another important potential source of metal ion contamination is the C18 material in the chromatographic columns employed in LC-ES-MS/MS. The data from the manufacturer's product information states that the total metal content is approximately 10 ng/mg (LC Packings product information). Since the silica packing material has been extensively washed during the manufacturing process, the metals can be considered as nonwashable or 'immobilized'.

The interaction of phosphopeptides and these immobilized metals could result in strong retention of the phosphopeptides such that they cannot be eluted by a common reversed phase gradient. This is the same mechanism employed by two popular enrichment procedures: immobilized metal affinity chromatography (IMAC) and titanium dioxide (TiO2) columns. In effect, we used a C18 column to enrich the multiple phosphorylated peptides from both - and - caseins and detected them with higher sensitivity and more selectivity.

These results provide some guidance for the scientists in the field of phosphopeptide analysis. Careful experiments using some standard peptides or model phosphoproteins can help to discover if one's LC system is contaminated by metal ions. Also, the addition of EDTA should also be done carefully and used only when it is necessary. In the three LC-ES-MS/MS systems employed, Fe was found as the main contaminant in two of these systems, with Ni in the third. Attention should also be paid to how much EDTA is introduced into their system, as it is related to the system design.

For optimal phosphopeptide analysis, an LC system with metal free components and a metal-free stationary phase would be the direction for the instrument development.

References
1.     Liu, S., Zhang C., Campbell, J.L., Zhang, H., Yeung K.K., Han, V.K.M., Lajoie, G.A. Rapid Commun Mass Spectrom, 2005. 19(19): 2747-56.
2.     Steen, H., Jebanathirajah, J.A., Rush J., Morrice, N., Kirschner, M.W. Mol Cell Proteomics, 2006. 5(1): 172-81.
3.     Larsen, M.R.; Graham, M.E.; Robinson, P.J.; Roepstorff, P. Molecular & Cellular Proteomics. 2004; 3: 456-65.
4.     Kim, J.; Camp II, D.G.; Smith, R.D. J. Mass Spectrom. 2004; 39: 208-15.


"Suya Liu, PhD, is currently a postdoctoral researcher fellow in Professor Lajoie's Biological Mass Spectrometry Laboratory at the University of Western Ontario, London, ON, Canada. His research interests are directed towards mass spectrometric analysis of protein, peptides and lipids."