The Human and Mouse Complement of SH2 Domain Proteins—Establishing the Boundaries of Phosphotyrosine Signaling
A comprehensive bioinformatic survey identified 120 SH2 domains across 110 human proteins, establishing structural and functional links to 18 human diseases and 81 mouse knockout phenotypes, with the authors emphasizing that this definitive catalog provides the necessary foundation for systems-level modeling of tyrosine kinase signaling networks rather than continued piecemeal analysis.
Dr. Monica Raina
Periodontist
Table of contents
Share
Background & context
The discovery that SH2 domains couple activated tyrosine kinases to downstream signaling pathways opened a productive line of research into phosphotyrosine-dependent protein interactions. Early structural and biochemical work — much of it from the same research group — established the general binding rules: SH2 domains recognize phosphotyrosine plus a handful of flanking residues, and they appear alongside tyrosine kinases at a specific point in eukaryotic evolution.
Yet this mechanistic understanding had outpaced a basic inventory problem: no comprehensive, non-redundant catalog of SH2 domain-containing proteins existed for any genome. Prior surveys disagreed on basic counts — one early genome-wide effort identified 87 SH2 proteins, a later one 98 — and suffered from duplication, incomplete annotation, and inconsistent curation. Without a definitive parts list, systems-level analysis of tyrosine kinase signaling rests on an uncertain foundation.
Establishing the scope of the SH2-ome
Earlier efforts treated cataloging as a counting exercise: how many SH2 domains exist, and in how many genes? That question, despite multiple attempts, had not converged on a stable answer.
A genome-wide, integrative approach resolves this by combining three complementary layers of analysis:
Sequence/genomic identification | Structural & disease annotation | Functional classification |
|---|---|---|
Bioinformatic identification (Pfam/SMART HMMs, BLAST, manual curation) of every non-redundant SH2 domain in human and mouse, removing duplicates, splice variants, and pseudogenes | Cross-referencing every known SH2 domain structure (169 structures across 43 domains) and every disease-associated mutation (18 proteins) or mouse knockout phenotype (81 genes) | Grouping the full set into 11 functional categories based on co-occurring domains (kinase, phosphatase, GEF, GAP, SH3, PTB, PH), mapping which downstream pathways each PTK signal can access |
This shift — from isolated domain-ligand characterization toward a complete, cross-referenced inventory — turns SH2 domains from a collection of individually interesting proteins into a bounded, systems-level resource: the full set of phosphotyrosine “readers” available to any activated tyrosine kinase.
Why a definitive catalog matters
For systems-level signaling analysis: Knowing the complete, non-redundant set of 120 SH2 domains in 110 human proteins (and an almost identical mouse complement) defines the maximum diversity of pTyr-dependent responses any single receptor can recruit — a necessary boundary condition for modeling kinase signaling networks rather than studying them one interaction at a time.
For human disease and mouse genetics: Mapping disease-causing mutations and knockout phenotypes onto the full SH2 catalog reveals that the same domain family produces strikingly different consequences depending on whether mutations cause loss of function (e.g., immunodeficiencies from SH2D1A or Btk mutations) or gain of function (e.g., the Shp2 mutations that cause Noonan syndrome by disrupting an autoinhibitory SH2–phosphatase interaction).
For drug discovery and evolutionary insight: A complete structural inventory exposes both well-studied subfamilies (with dozens of solved structures) and entirely uncharacterized ones, directing future structural and functional work toward genuine gaps rather than already-saturated domains — while also revealing unexpected biology, such as SH2 domains that bind phosphoinositides or carbohydrates rather than phosphotyrosine.
The open challenge
A static catalog, however comprehensive, is not the same as a working model of signaling. The harder task is connecting this parts list to actual cellular outputs — predicting, from kinase specificity and SH2 binding preference, which signaling complexes form under physiological conditions, and validating those predictions with quantitative methods like mass spectrometry and peptide/domain arrays.
Going forward
Establishing the definitive, structurally and functionally annotated set of SH2 domain proteins provides a benchmark against which systems-level studies of tyrosine kinase signaling can be built, rather than starting over with each new pathway studied. The same logic extends beyond SH2 domains: with roughly 100 distinct interaction domain families in the human proteome, comparable comprehensive catalogs for each family are a prerequisite for a genuinely systems-level understanding of how modular protein interactions generate cellular behavior — and for profiling how these signaling networks go wrong in disease.
