Phosphorylation Essentials¶
The Short Answer¶
Phosphorylation is the addition of a phosphate group to a protein, catalyzed by enzymes called kinases. It acts as a molecular switch that turns signaling pathways on or off. In cancer, kinases are among the most frequently dysregulated proteins, and kinase inhibitors represent one of the largest classes of targeted therapies. Understanding which sites are phosphorylated, by which kinase, and with what downstream effect is the central challenge of phosphoproteomics.
The Longer Answer¶
The signaling cascade¶
Phosphorylation operates through a hierarchical relay. An extracellular signal (e.g., growth factor) binds a receptor tyrosine kinase at the cell surface, triggering a cascade of kinase activations: receptor kinase phosphorylates an adaptor or intermediate kinase, which phosphorylates a downstream kinase, which ultimately phosphorylates transcription factors that alter gene expression. Each step amplifies and diversifies the signal. The reverse reaction is catalyzed by phosphatases, which remove phosphate groups and reset the switch. The balance between kinase and phosphatase activity determines the net phosphorylation state of any given site.
The signaling hierarchy¶
A simplified view of the relay:
- Receptor activation -- ligand binds receptor tyrosine kinase (e.g., EGFR, FGFR)
- Kinase cascade -- sequential activation of intracellular kinases (e.g., RAS-RAF-MEK-ERK)
- Effector phosphorylation -- terminal kinases phosphorylate transcription factors, metabolic enzymes, or structural proteins
- Cellular response -- changes in gene expression, proliferation, migration, or apoptosis
Key numbers¶
| Statistic | Value | Source |
|---|---|---|
| Human kinases (the "kinome") | 518 | Manning et al., Science 2002 |
| Known human phosphosites | ~240,000 | PhosphoSitePlus (2024) |
| Sites with known upstream kinase | <5% | Hornbeck et al., NAR 2015 |
| Sites with established function | <3% | Ochoa et al., Nature Biotechnology 2020 |
| Serine phosphorylation | ~86% | Olsen et al., Cell 2006 |
| Threonine phosphorylation | ~12% | Olsen et al., Cell 2006 |
| Tyrosine phosphorylation | ~2% | Olsen et al., Cell 2006 |
The disproportion between the number of detected sites and the number with functional annotation defines the annotation bottleneck. Mass spectrometry can measure phosphosites at scale, but biological interpretation lags far behind.
Why Phosphoproteomics Matters for Cancer¶
Cancer hijacks normal signaling by activating kinases that drive proliferation and survival. Three classes of oncogenic kinase dysregulation dominate the therapeutic landscape:
- Gain-of-function mutations -- BRAF V600E constitutively activates the MAPK pathway; targeted by vemurafenib and dabrafenib.
- Amplification or overexpression -- HER2 (ERBB2) amplification in breast cancer; targeted by trastuzumab and lapatinib.
- Fusion proteins -- BCR-ABL in chronic myeloid leukemia; targeted by imatinib.
Kinase inhibitors account for roughly one-third of all FDA-approved targeted cancer therapies. Yet resistance emerges in nearly all cases, frequently through bypass signaling: when one kinase is inhibited, parallel or downstream kinases compensate. Phosphoproteomics is uniquely suited to detect these adaptive rewiring events because it measures the functional output of kinase activity directly, rather than relying on genomic proxies.
Why Prediction Is Hard¶
Assigning a phosphosite to its upstream kinase is not a simple lookup. Multiple factors make computational prediction unreliable without experimental validation.
| Challenge | Explanation |
|---|---|
| Motif degeneracy | Many kinases recognize overlapping sequence motifs; a basophilic motif (R-x-x-S) is shared by AKT, PKA, PKC, and others |
| Context-dependent specificity | Subcellular localization, scaffolding proteins, and tissue-specific expression constrain which kinase-substrate pairs interact in vivo |
| Stoichiometry unknowns | A site may be detected at low occupancy; it is unclear whether this reflects regulatory phosphorylation or background noise |
| Tissue-specific kinase expression | A kinase expressed in brain may not be relevant to a phosphosite detected in lung tumor tissue |
| Sparse ground truth | Training data for prediction tools comes from the same small set of validated kinase-substrate pairs, creating circular dependencies |
| Dynamic range | Phosphorylation changes on the timescale of seconds to minutes, but most proteomics experiments capture a single snapshot |