Switch-like cyclin-dependent kinase (CDK)-1 activation is thought to underlie the abruptness of mitotic onset, but how CDKs can simultaneously phosphorylate many diverse substrates is unknown, and direct evidence for such phosphorylation dynamics in vivo is lacking. Here, we analysed protein phosphorylation states in single Xenopus embryos throughout synchronous cell cycles. 22% of 4583 phosphosites on 1843 proteins were dynamic in vivo. We assigned cell cycle phases using egg extracts, showing 693 S-phase and 1035 mitotic phosphorylations. High-time resolution targeted phosphoproteomics in single embryos revealed switch-like mitotic phosphorylation of diverse protein complexes. 60% of cell cycle-regulated phosphosites occurred in CDK consensus motifs, and 72% located to intrinsically disordered regions (IDRs). Dynamically phosphorylated proteins, and documented CDK substrates in human and yeast, are significantly more disordered than targets of other cell cycle kinases and phosphoproteins in general. Furthermore, 30-50% are components of membraneless organelles. Our results suggest that CDK-mediated phosphorylation of intrinsic disorder allows switch-like mitotic cellular reorganisation.

X. laevis single embryos were collected individually and snap frozen. Embryos were lysed and protein digestion was performed using the FASP method. Phosphopeptides were enriched using Fe(III)-NTA 5 μL cartridges in the automated AssayMAP Bravo Platform (Agilent Technologies). Measurements were done on an EASY-nLC 1200 System (Thermo Fisher Scientific) coupled to an Orbitrap Q Exactive HF. Peptides were separated using an EASY-Spray analytical column (ID of 75 μm and 25 cm length; packed with 2 μm C18 particles with a 100 Å pore size) (Thermo Fisher Scientific). Gradient length was 60 minutes. Synthetic standards were used during method development for retention time scheduling and instrument ion fill-time optimization. Additionally, synthetic heavy peptides were pooled and combined with synthetic retention time peptide standards (iRT, Biognosys) to generate a spectral library, measured in DDA mode using the same LC-MS setup. This spectral library provided fragment intensity and retention time information for quality control assessment of targeted measurements. Samples were reconstituted in 2% FA containing ~200 fmol of each synthetic standard. The mass spectrometer was operated in data independent acquisition mode with an inclusion list of targets for parallel reaction monitoring (PRM). The list included the m/z values for the heavy and light versions of the phosphopeptides. Optimal measurement parameters were determined using test samples spiked with the heavy-labeled standards in order to guarantee optimal sensitivity for detection of endogenous phosphopeptides. We measured the targets of interest in a scheduled fashion, during a four-minute window with a 120,000 resolution, maximum injection time of 246 ms and an accumulation target value of 2e5 ions, in order to ensure maximum specificity and sensitivity.

X. laevis single embryos