Phosphorylated α-synuclein at Ser¹²⁹ (PS-129) was detected in human CSF using linear and tandem MS. (A to C) Linear mode spectra of intact proteins are shown for (A) PS-129 standard (0.5 ng) at a mass of 14,540; (B) candidate PS-129 in CSF after IP with a rabbit anti-PS-129 antibody; and (C) absence of PS-129 after IP with normal rabbit immunoglobulin G. (D) The digestion of the eluate from IP (with anti-PS-129 antibody) revealed two peptides with PS-129 phosphorylation (see also fig. S5). Phosphorylation of Ser¹²⁹ was also validated by manual mass match. The dethiomethyl modification of Met¹²⁷ is likely a sample processing artifact, and the loss of N-terminal Asp in the peptide shown is probably due to the acid lability of the Asp-Pro bond, both of which are known artifacts in proteomics analyses. (E) The Luminex assay specificity was confirmed by several methods including using pre-depletion of PS-129 from CSF with an anti-PS-129 antibody, de-phosphorylation (DP) of CSF proteins using calf intestine alkaline phosphatase, non-phosphorylated α-synuclein standard (α-Syn), and a non-specific phosphorylated protein β-casein as negative controls. BG, background.

(A to F) PS-129 concentrations in CSF were measured in the discovery set [(A) PS-129, (B) total α-synuclein, (C) ratio of PS-129 to α-synuclein] and validation set [(D) PS-129, (E) total α-synuclein, (F) ratio of PS-129 to α-synuclein). Subjects younger than 50 years were excluded. Comparisons of the average concentration of PS-129 in the CSF of healthy controls, PD patients, MSA patients, PSP patients, and AD patients were performed using one-way ANOVA. Case numbers included in each group can be found in table S1. Data are mean ± S.E. *, P<0.05; **, P<0.01; ***, P<0.001 (Tukey/ANOVA). Only the most relevant statistical significance is indicated; see table S1 for a complete report of all group comparisons.

PS-129 and total α-synuclein concentrations in CSF were measured by Luminex assays in the discovery and validation patient cohorts. ROC analysis was performed to evaluate the diagnostic values (sensitivity and specificity) of PS-129 alone, total α-synuclein alone, and the combination of total α-synuclein and PS-129 (in a logistic regression model) for PD diagnosis (not shown) and differential diagnosis between PD and other parkinsonian disorders (MSA and PSP). (A to C) Discovery set: (A) PD versus MSA, (B) PD versus PSP, (C) MSA versus PSP. (D to F) Validation set: (D) PD versus MSA, (E) PD versus PSP, (F) MSA versus PSP. Subjects younger than 50 were excluded from all analyses and samples with high blood contamination (hemoglobin ≥200 ng/ml) were excluded when total α-synuclein was measured. The diagnostic values in the discovery set when specificities or sensitivities were anchored to be ≥80% can be found in tables S2 and S3, respectively. The diagnostic values for the combination in the validation set when the models and cutoff values generated from the discovery set were applied into the validation set can be found in table S4.

PS-129 concentrations in the CSF of patients with PD were measured by a Luminex assay and plotted against PD severity measured by UPDRS motor scores. (A and B) The correlation between CSF PS-129 concentrations and UPDRS scores was evaluated using linear regression analysis with Pearson’s correlation in the discovery set [(A) r=0.23, P<0.05], the validation set [(B) r=0.21, p<0.05], and the two sets combined (not shown; r=0.22, P<0.01).