Identification of LSP1 Phosphorylation Sites

The phosphorylation sites of LSP1 were identified from samples resolved by 2-D gel electrophoresis. The LSP1 spots were excised from preparative scale 2-D gels and cleaved into peptides by in-gel digestion. Phosphorylated peptides from each spot were detected by nanospray precursor ion scanning as described in Methods and Protocols. Sample data for two WEHI-231 cell spots (L4 and L6) are shown in Fig. 6. The full mass spectrum, performed in negative ion mode, of a tryptic digest of spot L6 revealed a number of predominant peaks (Fig. 6A). Precursor ion scanning was then used to detect the m/z -79 (PO^3-) product ions released during MS/MS. A large number of phosphorylated peptide ions were detected in spot L6 (Fig. 6B). Most of the phosphopeptides were either minor components or undetectable in the full mass spectrum. The phosphorylated peptides were all present as multiply charged species, leading to the plethora of detectable peaks. A total of four phosphorylated peptide ions were detected (labeled T1, T2, T4, and T5 in Fig. 6B). As expected, the full mass spectrum of the tryptic peptides derived from spot L4 (a more highly phosphorylated species) was essentially identical to that of spot L6 (compare Fig. 6C and 6A). Precursor ion scanning of spot L4 detected two peptides (T1 and T2) that were also detected in spot L6 (Fig. 6D). Phosphopeptide T3, which was not detected in spot L6 (Fig. 6B), was a predominant phosphopeptide in spot L4 (Fig. 6D). In contrast, phosphopeptides T4 and T5 appeared to be absent in spot L4. As shown below, peptides T3, T4, and T5 are differentially phosphorylated forms of the same tryptic peptide (Table 1). Spot L4 contains the quadruply phosphorylated form of the peptide (T3), and spot L6 contains the doubly (T5) and triply (T4) phosphorylated forms of the peptide. These observations suggest that an important component in the acidic shift of spot L4 relative to L6 is due to phosphorylation of additional sites within this tryptic peptide. In order to cover more of the LSP1 sequence and identify additional phosphorylation sites, the LSP1 spots were also digested with Glu-C protease. The results of Glu-C digestion of spot L6 are shown in Fig. 6E and 6F. As observed following tryptic digestion, the phosphorylated peptides detected by precursor ion scanning were not readily detectable in the full mass spectrum (Fig. 6E and 6F). A total of five phosphorylated Glu-C peptides were detected in spot L6. These represented a total of four distinct peptides. Peptides G3 and G4 differed by a single amino acid due to cleavage at adjacent glutamic acid residues (Table 1).

Fig. 6. Identification of phosphorylated LSP1 peptides by precursor ion scanning. WEHI-231 LSP1 was resolved by preparative 2-D gel electrophoresis. Spots L4 and L6 of LSP1 were excised from 2-D gels and digested with either trypsin or Glu-C protease. The resulting peptide mixtures were analyzed by precursor ion scanning using nanospray mass spectrometry in negative ion mode. (A) Full mass spectrum of the tryptic digest of spot L6. (B) Precursor ion scan of the tryptic digest of spot L6. (C) Full mass spectrum of the tryptic digest of spot L4. (D) Precursor ion scan of the tryptic digest of spot L4. (E) Full mass spectrum of the Glu-C digest of spot L6. (F) Precursor ion scan of the Glu-C digest of spot L6; due to limited amounts of sample, MS data were only collected for m/z 600 to 1200. In panels A, C, and E, the major LSP1 peptide ions are labeled with their m/z values. In panels B, D, and F, the major phosphate-containing peptide ions detected in the precursor scan are labeled with a number corresponding to the peptide from which the ions were derived (T1 to T5 are tryptic peptides; G1 to G5 are Glu-C peptides). The charge states (2- to 6-) of the peptide ions detected in the precursor scans are indicated by superscript numbers.

Ten phosphorylated peptide ions were detected by precursor ion scanning of tryptic and Glu-C peptides derived from LSP1. This represented a total of six distinct peptides after taking differential modifications into account. The representative data shown in Fig. 6 were derived from LSP1 spots L4 and L6 from WEHI-231 cells. Precursor ion scanning was also performed on spots L5 and L7 from both WEHI-231 and B cells and spots L4 and L6 from B cells. While the numbers of phosphopeptides detected varied between spots, the same phosphopeptides found in spot L4 were also found in the other spots. No additional phosphopeptides were detected, indicating that each spot contained the same set of phosphopeptides. Fewer phosphopeptides were detected in spots L6 and L7, which would be consistent with a lower level of phosphorylation. However, the nonquantitative nature of the mass spectrometry analysis precludes making any conclusions about whether sites in LSP1 are phosphorylated in a random or a hierarchical manner.

After identification of phosphorylated peptides by precursor ion scanning, individual phosphorylation sites were identified by nanospray tandem mass spectrometry. Once the m/z ratios of phosphopeptides were determined by precursor ion scanning in negative ion mode, the peptide (precursor) ions at the corresponding m/z values in positive ion mode were fragmented by collision-induced dissociation. The resulting MS of the product ions provided partial sequence information that was used to identify the sites of phosphorylation. In most cases, a signal for the precursor ions was not visible in the positive mode MS scan. Nevertheless, fragmentation at the appropriate m/z value gave clear product ion spectra derived from the correct peptide. Whenever possible, tandem mass spectrometry was performed on both the doubly and triply charged forms of the precursor ions. Product ion spectra for tryptic peptide T2 from WEHI-231 spot L4, Glu-C peptide G1 from WEHI-231 spot L4, and Glu-C peptide G2 from B-cell spot L4 are shown in Fig. 7. The partial sequence data obtained allowed assignment of four phosphorylation sites in these three peptides (Table 1). Peptides T1 and T2 were identical except for the presence of oxidized methionine in T2 (a common artifact seen in mass spectrometry). Tandem mass spectrometry of peptide T1 identified Ser-243 as the phosphorylation site in this tryptic peptide (not shown). Phosphorylation sites at Ser-180, Ser-184, and Ser-205 were identified from the product ion spectra of Glu-C peptides G1 and G2. As discussed above, peptides T3 through T5 were differentially phosphorylated forms of the same peptide (amino acids 166-188). As summarized in Table 1, the product ion spectra of these peptides indicated a total of four phosphorylation sites (data not shown). However, the spectra contained insufficient information to assign these sites due to the low abundance of the precursor ions. The product ion spectrum of Glu-C peptide G2, which overlapped with the C-terminal region of the 166-188 tryptic peptide, allowed identification of Ser-180 and Ser-184 as two of the four sites phosphorylated. The product ion spectra of Glu-C peptide G3/G4, which overlap with the amino terminal region of the 166-188 tryptic peptide, both contained two phosphates that could not be assigned. The G3/G4 peptide (157-EHLIRHQVRTPSPLALEDTVE-177) contains two threonines and one serine (indicated by asterisks in Fig. 8). The product ion spectrum of peptide G5 identified an additional phosphorylation site in an amino terminal LSP1 peptide (66-QPGQQTLISLKSSE-82). Although precursor ions from this peptide were analyzed multiple times, unambiguous assignment of the phosphorylation site was not possible. The four potential phosphorylation sites in this peptide are indicated by asterisks in Fig. 8.

Fig. 7. Identification of LSP1 phosphorylation sites by tandem mass spectrometry. Proteolytic digests of LSP1 were prepared as described for Fig. 6 and analyzed by tandem mass spectrometry. Following identification of phosphorylated peptides by precursor ion scanning in negative ion mode, peptide ions at the corresponding m/z values were analyzed by tandem mass spectrometry in positive ion mode. The MS of the product ions resulting from collision-induced dissociation of LSP1 peptides is shown. The product ion peaks (either b series or y series) that were most diagnostic of the phosphorylation sites are indicated along with the sequence of the peptide. The fragments representing individual b or y ions are indicated by lines between the corresponding amino acids. The charge states of the product ions are indicated in superscript. Ions with no superscript numbers are singly charged. In several cases, peaks corresponding to product ions containing phosphate (bn or yn), and satellite peaks resulting from neutral loss (98 Da) of phosphoric acid (bn-H3PO4 or yn-H3PO4) were detected. (A) Product ion spectrum of the doubly charged tryptic peptide T2 precursor ion at m/z 819.9 from WEHI-231. In addition to neutral loss of phosphoric acid, all of the product ions labeled in this spectrum had masses that were 17 Da lower than the actual mass due to the neutral loss of [NH3]. (B) Product ion spectrum of the triply charged Glu-C peptide G1 precursor ion from WEHI-231at m/z 984.2. In addition to product ions, the mass spectrum also contains a peak corresponding to the triply charged precursor ion (M3+) and a satellite peak resulting from neutral loss of phosphoric acid at m/z 935.2 (M-H3PO4)3+. (C) Product ion spectrum of the doubly charged Glu-C peptide G2 precursor ion at m/z 987.0 from B cells. This spectrum also contains a peak corresponding to the precursor ion (M2+) and a peak resulting from neutral loss of phosphoric acid (M-H3PO4)2+

The collection of 10 phosphorylated peptides from LSP1 identified a total of 7 sites of phosphorylation. Four of these were assigned to Ser-180, Ser-184, Ser-205, and Ser-243 (Fig. 8). Two of the remaining sites were localized to a Glu-C peptide containing Thr-166, Ser-168, and Thr-175. The final site was localized to a Glu-C peptide containing Thr-71, Ser-74, Ser-77, and Ser-78. The presence of seven phosphorylation sites is consistent with the resolution of LSP1 into seven charge forms by 2-D electrophoresis. The fact that these sites were all identified in resting cells indicates that LSP1 is highly phosphorylated under basal conditions. A total of 64% of the LSP1 sequence was covered by the tryptic and Glu-C peptides identified in these experiments. A significant number of serine, threonine, and tyrosine residues are present in peptides that were not detected, and LSP1 could contain additional sites of phosphorylation.

Fig. 8. Summary of the phosphorylation sites identified in LSP1. The amino acid sequence of the long splice variant of murine LSP1 is shown. Orange underlining indicates the tryptic peptides that were identified. Blue underlining indicates the Glu-C peptides identified. Residues identified as phosphorylation sites are in boldfaced type and marked with the amino acid residue number. Possible phosphorylated residues within peptides where the actual sites could not be identified are marked by asterisks. These included a single phosphorylation site within peptide G5 (residues 66-82) and two phosphorylation sites within peptide G3/G4 (residues 157-177).

To determine if the phosphorylation sites identified here might be substrates for known protein kinases, predicted sites of phosphorylation of LSP1 were identified by analyzing the LSP1 amino acid sequence with NetPhos , PROSITE, and Scansite phosphorylation site prediction programs. The results of these searches are shown in Table 2. All four of the directly identified sites (Ser-180, Ser-184, Ser-205, and Ser-243) were predicted to be sites of phosphorylation by both NetPhos and Scansite. While PROSITE predicted Ser-205 to be a PKC site, the other sites were not predicted to be phosphorylated. This is likely due to the limited number of kinase phosphorylation site motifs included in PROSITE. Ser-205 was predicted to be a PKC site by both PROSITE and Scansite and is very likely to be an in vivo site of PKC phosphorylation. Ser-243 had a very low Scansite percentile score as a site for calmodulin-dependent protein kinase 2 (CaMKII) indicating a close match to the optimal sequence. This residue was previously proposed to be a MAPKAP kinase 2 phosphorylation site(12). While Ser-184 was a predicted site for several proline-directed kinases, only ERK1 had a Scansite percentile score below the standard cutoff of 0.2(15). Of the putative kinase motifs containing Ser-180, only GSK3 had a Scansite percentile score below 0.2. There were no obvious candidate protein kinases for the phosphorylation sites present in peptides G3/G4 or G5. Of the three possible phosphorylation sites on Glu-C peptide G3/G4, only Ser-168 was predicted as a phosphorylation site by NetPhos. Of the four possible sites in peptide G5, Ser 78 had the highest NetPhos score. While it is not clear how accurate these predictions will be, they suggest that LSP1 is likely to be a substrate of multiple protein kinases including CaMKII, GSK3, MAPKAP kinase 2, PKC, and ERK1.