Supplementary document 1

Definition of site-determining ion combination (SIDIC)


A mono-phosphorylated peptide, XXXXSXXSXXYXX, has three candidates in terms of S/T/Y phosphorylation site (S5, S8 and Y11) (Figure SD1-1). In order to indicate S5-phosphorylation exclusively, at least one ion in the yellow box (phospho-b5, phospho-b6, phospho-b7, nonphospho-y6, nonphospho-y7 and nonphospho-y8) must be observed (candidate (A)). Similarly, at least one ion in the green box (nonphospho-b8, nonphospho-b9, nonphospho-b10, phospho-y3, phospho-y4 and phospho-y5) must be found in MSMS spectrum to support Y11-phosphorylation exclusively (candidate (C)). On the other hand, to prove S8-phosphorylation exclusively, at least one ion each in two boxes (yellow box: nonphospho-b5, nonphospho-b6, nonphospho-b7, phospho-y6, phospho-y7 and phospho-y8, and green box: phospho-b8, phospho-b9, phospho-b10, nonphospho-y3, nonphospho-y4 and nonphospho-y5) is required (candidate (B)), i. e., it is necessary to observe at least one fragment ion in each side. We call this ion relationship ‘site-determining ion combination (SIDIC)’. The existence of observed SIDIC can prove the localization of the phosphorylation site exclusively.

It is possible to generalize how to find SDIC based on the examples above. When you have phosphorylated residue at position X and other phosphorylation candidate positions are V, W, Y and Z (V<W<X<Y<Z from N-terminus) in peptide with 10 residues, the fragment ions you have to pay attention are divided into two groups. Group 1 consists of non-phosphorylated y-ions of 11-Y to 10-X and phosphorylated b-ions of X to Y-1, and Group 2 consists of phosphorylated y-ions of 11-X to 10-W and non-phosphorylated b-ions of W to X-1. SIDIC consists of at least one ion in each group. The number of SIDIC is defined as the number of ions in Group 1 by that in Group 2. Note that fragment ions with the same m/z values (less than 0.8 m/z difference) as other b,y-ions with or without neutral loss of H3PO4 are not considered for SIDIC evaluation.

Determination of phosphorylation site localization based on SIDIC

Figure SD1-2 shows the flowchart of site-determination using SIDIC. Depending on the number of phosphorylated sites, candidates for modification variants are generated. If there is no variant (i. e., only one STY residue exists for monophosphorylated peptide for instance), the identified result is determined as ‘unambiguous’. For each candidate, the number of observed and observable SIDIC is calculated. Peaks from neutral loss of phospho-group are not considered.


As a first filter, probability-based score of each modification variant is calculated using the same manner as PTM score (Olsen et al., 2006), and candidates with delta score >20 are rejected, whereas candidates with top score are accepted if at least one SIDIC is observed. For other candidates, if the ratio of observed to observable SIDIC is over 0.3, the candidates are accepted as possible hits. If only one possible hit is accepted from one MS/MS spectrum, the final result is indicated as ‘unambiguous’. If more than one possible hit are survived, the MS/MS spectrum may consist of multiple modification variants and the conclusion is shown as ‘indistinguishable’. If no possible hit is accepted, the conclusion is described as ‘undetermined’.

Examples of SIDIC-based determination of phosphorylation site localization

An MS/MS spectrum from the rice sample was identified as monophosphorylated VHESPVLSPQR. Along the flowchart shown in Figure DS1-2, two candidates were survived as shown in Figure DS2-3(A) and 3(B) (Supplementary Table 1, Rice SIDIC sheet, No. 6528). Annotated peaks in spectra (A) and (B) supported the mono-phosphorylation at S4 and S8, respectively. The fragment ion y4 in spectrum (B) was not used for SIDIC since it was not distinguishable from doubly-charged ions of y10-98. As a result, the SIDIC ratio of observed to observable ions was 4 / 8 for the peptide with S4-phosphorylation, and 6 / 8 for S8. Based on it, it was concluded that the MS/MS spectrum might contain fragments derived from both modification variants and the phosphorylation site localization was decided to be ‘indistinguishable’. Note that di-phosphorylated VHEpSPVLpSPQR was also identified and the localization of both phosphosites was confirmed as ‘unambiguous’ (Supplementary Table 1, Rice SDIC sheet, No. 6529).


Reference

Olsen JV, Blagoev B, Gnad F, Macek B, Kumar C, Mortensen P, Mann M (2006) Global, in vivo, and site-specific phosphorylation dynamics in signaling networks. Cell 127: 635-648