Six new autophosphorylation sites were detected in the catalytic subunit of protein kinase A using 1D-PAGE, in-gel digestion and UPLC-MS/MS. In addition, the degree of phosphorylation at these sites was determined by a standard-free, signal intensity-based method with the following characteristics, (i) the use of three proteases, (ii) consideration of the interplay between phosphorylation and proteolysis, and (iii) the addition of citrate to achieve unbiased recovery of phosphopeptides in UPLC.

Results

PKA is the most important model for kinases and has been extensively studied (reviewed in [1]). Four autophosphorylation sites in the PKA catalytic subunit α have been previously reported, namely pS10, pS139, pT197, and pS338. In this study six new autophosphorylation sites were identified using UPLC-MS/MS, namely pS14, pT48, pS53, pS212, pS259, and pS325 (fig. 1). In addition, a standard-free method for estimation of the site-specific degree of phosphorylation was developed.

A bottleneck in analyzing reversible protein phosphorylation by LC-ESI-MS/MS is that frequently low and erratic recoveries from RP-columns of phosphopeptides are observed, especially in connection with the analysis of multiply phosphorylated peptides. This drawback hampers the direct deduction of the site-specific degree of phosphorylation from uncorrected signal intensities. We suspected, that irreversible adsorption of metal-ions on reversed phase columns (fig. 2) is a major mechanism responsible for the incomplete recovery of phosphopeptides. To destroy the three component system outlined in figure 2 and thus increase the phosphopeptide recovery, the addition of metal ion-mobilizing ligands to the samples was demonstrated to be a promising strategy [2]. The acronym mimLC (for metal ion-mobilizing LC) was introduced for LC performed with addition of metal ion-complexing sample additives [3]. In mimLC, the complete LC system including the stationary phase is depleted from metal ions. In this way, metal ion-mediated adsorption of polar analytes such as phosphopeptides is minimized since adsorbed metal ions are eluted as metal ion complexes. In conclusion, we found three metal ion-mobilizing additives to significantly increase the recovery of phosphorylated and multiphosphorylated peptides, especially at subpicomole amounts, namely citrate, EDTA and the tetrapeptide pSpSpSpS.

All analyses of PKA were hence performed with support of citrate as metal ion-mobilizing additive.

Using mimLC, signal intensities of phosphorylated and unmodified peptides can be compared directly. For the set-up of the standard-free approach, all signal intensities of phospho-, and non-phosphopeptides containing a particular site that were observed in the different digestion experiments were considered comprehensively. Furthermore, all charge states were included in the evaluation (fig. 3). The data show that the 10 PKA Cα autophosphorylation sites fall into four distinct groups which correlate the measured phosphorylation degrees with the motifs around the phosphorylated sites (fig.4). Known phosphorylation sites were almost completely phosphorylated. New phosphorylation sites in a PKA consensus motif (R/K-X-X-pS/T) are phosphorylated at approximately 25%, phosphorylation sites located in an inverted PKA consensus motif are phosphorylated at a lower level, and sites in motifs without similarity to the PKA consensus motif are only marginally phosphorylated.

Discussion

In summary, estimation of phosphorylation degrees based on uncorrected ion intensity ratios (standard-free method) will provide data with slight systematic underestimation [4]. Nevertheless, the data obtained show a high degree of internal consistency, proving that accurate data may be obtained even without the use of isotopically labeled standards, which have to be generated individually for each site. Recently, this standard-free method has been used in an in-depth analysis of the kinase/substrate system PfCDPK1/PfGAP45 [5]. In this way it was observed e.g. that two closely adjacent sites are phosphorylated in a mutually exclusive manner. This straightforward approach may also be extended to the relative quantification of other post-translational modifications.

Methods

His-tagged wild-type PKA Cα was expressed in E. coli BL21(DE3) and purified using Ni-IMAC and 1D-PAGE. In-gel digestion by trypsin, AspN and chymotrypsin were done as described [7]. UPLC-MS/MS was done on a Waters system (nanoAcquity, QTOF2) with addition of 50 mM citrate [2].

References
[1] Gesellchen F., et al., Biochim Biophys Acta - Proteins & Proteomics 1764, 1788-1800 (2006)
[2] Winter D., et al., J Proteome Res 8, 418-424 (2009)
[3] Seidler J., et al., Amino Acids, DOI 10.1007/s00726-010-0647-7 (2010)
[4] Steen H., et al., Mol Cell Proteomics 5, 172-181 (2006)
[5] Winter D., et al., Anal Biochem 393, 41-47 (2009)
[6] Seidler J., et al., Anal Bionanal Chem 395, 1713-1720 (2009)
[7] Shevchenko A., et al., Nat Protoc 1, 2856-2860 (2006)