⚛️ biophysics

Conservation and divergence in the allosteric architectures of five human protein kinases

By mapping over 160,000 variants across five human kinases, this study reveals that while some allosteric sites are conserved, each kinase possesses a unique and divergent energetic architecture with distinct functional secondary sites, offering new opportunities for developing specific therapeutic inhibitors and activators.

Original authors: Folgado, C., Beltran, A., Lehner, B.

Published 2026-08-07

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Folgado, C., Beltran, A., Lehner, B.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine your body is a bustling city, and inside every cell, there are thousands of tiny workers called proteins. Among these, a special group called kinases act as the city's master switches. Their job is to flip other proteins "on" or "off" by attaching a tiny chemical tag to them, much like a foreman handing out a "Go" or "Stop" sign. These switches control everything from how your cells grow to how they repair damage. However, when these switches get stuck in the "on" position, it can lead to serious trouble, like cancer.

To fix broken switches, scientists often try to build "brakes" (drugs) that jam the main control panel, known as the active site. But here's the problem: almost all kinases look incredibly similar at this main control panel. It's like trying to jam a specific lock on a thousand different doors when they all use the exact same keyhole. If you jam one, you often jam them all, which can cause nasty side effects because you're stopping the good workers too. Scientists have long wondered if these switches have hidden, secondary control panels (called allosteric sites) that are unique to each type of switch. If they do, we could build super-specific brakes that only stop the bad switch without touching the others. But for most of these proteins, we didn't know where these hidden panels were, or if they even existed.

This paper takes a giant leap forward by mapping the "energy landscape" of five different human kinases: SRC, FGR, JNK2, ZAK, and TSSK2. The researchers used a high-tech method called KINASE-MAPS to test over 160,000 tiny mutations (changes) in these proteins. Imagine taking a giant Lego set of a protein and swapping out every single brick for every other color of brick, one by one, to see how it changes the machine's behavior. They measured two things for every single change: did the protein fall apart (abundance), and did it stop working or start working too hard (activity)?

The results revealed a fascinating mix of similarity and surprise. First, the team confirmed that for stopping a kinase, the rules are somewhat predictable: the closer you are to the main control panel, the more likely a change is to break the machine. This "distance-dependent" rule is like a shockwave; if you hit the center, the whole thing shakes. However, the direction of that shockwave is different for every kinase. Some kinases are more sensitive to hits from the top, others from the side. This means the "shape" of the shockwave is unique to each protein.

Even more exciting is what they found about turning kinases on. While stopping them follows a general rule, activating them is a completely different story. The paper shows that the "on" switches are scattered all over the protein, and they are totally different for each kinase. A spot that acts as a secret "on" button for one kinase might be a "do nothing" spot for another, or even a "break" spot for a third.

The researchers also mapped out 37 different "pockets" or nooks on the surface of these proteins where a drug could potentially stick. They found that while a few pockets are shared by all five kinases (making them good targets for broad-spectrum drugs), most pockets are unique. Some pockets are "switchable": a drug that turns one kinase on might accidentally turn another one off. This proves that even though these proteins look like twins from the outside, their internal wiring and hidden control panels are completely unique.

In short, the paper suggests that there is no "one-size-fits-all" map for these proteins. To build the perfect, non-toxic medicine, we can't just guess where the hidden controls are based on how one kinase looks. Instead, we need to build a custom map for every single protein we want to target. This discovery opens the door to designing drugs that are so specific they can stop a bad cell without ever bothering its healthy neighbors, turning the tide in the fight against diseases like cancer.

Technical Summary: Conservation and Divergence in the Allosteric Architectures of Five Human Protein Kinases

Problem Statement
Protein kinases are central to biological regulation and are frequent targets for therapeutic intervention, with nearly 100 clinically approved drugs. However, the development of specific kinase inhibitors is hindered by the high structural conservation of the ATP-binding active site, leading to off-target toxicity and resistance mutations. While targeting functional secondary (allosteric) sites offers a path to increased specificity and the ability to activate kinases, the functional secondary sites for the vast majority of kinases remain unknown. Furthermore, it is unclear whether the energetic and allosteric architectures of kinases sharing the same structural fold are highly conserved or if they diverge significantly, resulting in protein-specific regulatory networks.

Methodology
The authors employed a high-throughput experimental platform termed KINASE-MAPS (Kinase-Mutational Allosteric Propagation Scan) to generate complete energetic and allosteric maps for five human protein kinases: SRC, FGR, JNK2 (MAPK9), ZAK (MAP3K20), and TSSK2. These kinases were selected to represent diverse physiological roles, disease associations, and evolutionary distances (two SRC-family tyrosine kinases and three serine/threonine kinases from distinct groups).

The methodology involved four primary steps:

Library Construction: Using SUNi mutagenesis, the authors constructed combinatorial double-mutant libraries covering the entire kinase domains (269–357 residues) of each protein. These libraries encoded all possible single amino acid substitutions within eight distinct genetic backgrounds to constrain model fitting.
Activity Quantification: A cell-based toxicity assay in Saccharomyces cerevisiae was used to quantify kinase activity. Toxicity was validated as activity-dependent, and selections were performed across three independent replicates.
Abundance Quantification: A growth-based selection using a sandwich protein-fragment complementation assay (sPCA) quantified the folded protein abundance of every library variant to distinguish effects on catalysis from effects on protein stability.
Thermodynamic Modeling: A phenomenological energy model (MoCHI) was fitted to the combined activity and abundance data. This model assumes three protein states (unfolded, folded-inactive, folded-active) and infers the change in folding free energy ( $\Delta\Delta G_f$ ) and activity free energy ( $\Delta\Delta G_a$ ) for each single amino acid variant, assuming additivity in double mutants.

Key Contributions and Results
The study quantified the effects of over 160,000 variants, generating a dataset of ~50,000 free energy changes across the five kinases. The key findings include:

Conservation of Fold Stability vs. Divergence of Allosteric Architecture: While destabilizing mutations are generally enriched in the protein core across all five kinases, the specific structural regions critical for stability vary. For instance, destabilizing mutations are enriched in the C-lobe for SRC, FGR, and ZAK, but in the N-lobe for JNK2.
Distance-Dependent but Anisotropic Inhibition: Inhibitory allosteric mutations exhibit a conserved distance-dependent decay from the active site across all five kinases, with a median half-distance ( $d_{1/2}$ ) of 9.3 Å. However, this allostery is highly anisotropic; coupling is strongest toward the $\alpha$ C-helix and weakest toward the $\alpha$ D-helix. The strength of this coupling varies significantly between kinases in specific directions (e.g., JNK2 shows strong coupling toward the activation loop, whereas ZAK does not).
Divergence of Activation Sites: Unlike inhibitory mutations, activating mutations are not enriched near the active site. Instead, they are structurally dispersed and highly kinase-specific. Only a small fraction of activating sites are shared between kinases, and the structural distribution of these sites differs markedly (e.g., enriched in the C-lobe for SRC/JNK2/ZAK but the N-lobe for TSSK2).
Conserved and Kinase-Specific Allosteric Pockets: Using FTMap, the authors identified 37 distinct fragment-sized pockets across the five kinases.
- Conserved Inhibitory Pockets: Five pockets (including the DFG, MT3, and two uncatalogued inter-lobe pockets) are enriched for inhibitory mutations in all five kinases, suggesting a functionally conserved inhibitory surface.
- Kinase-Specific Pockets: Each kinase possesses a unique set of additional inhibitory and activating pockets. For example, JNK2 has specific pockets corresponding to its substrate-recruitment surfaces (DRS and DEF), while SRC has unique pockets near the $\alpha$ C helix.
- Switching Pockets: Several pockets exhibit "switching" behavior, where mutations are activating in one kinase but inhibitory in another (e.g., the PIF pocket is activating in JNK2 but inhibitory in TSSK2).
Surface Divergence: While the two most closely related kinases (SRC and FGR) share similar surface mutational effects, more distant kinases show substantial divergence in their allosteric surfaces.

Significance
The paper claims that these results provide the first comprehensive comparative analysis of the energetic and allosteric structures of homologous proteins within an enzyme family. The authors conclude that while kinases share a conserved catalytic core and certain principles of allosteric decay (distance-dependent, anisotropic), their allosteric architectures are unique to each protein.

The significance of this work lies in demonstrating that:

Protein-Specificity is the Rule: Despite sharing the same 3D fold, each kinase has a distinct set of functional secondary sites for regulation and therapeutic targeting.
Therapeutic Implications: The identification of conserved inhibitory pockets offers potential targets for broad-spectrum modulation, while the discovery of kinase-specific pockets (both inhibitory and activating) provides a rational basis for developing highly specific drugs to overcome the selectivity and resistance issues associated with ATP-competitive inhibitors.
Evolutionary Insight: The data suggests a general principle in protein evolution where a conserved catalytic core is overlaid with divergent regulatory control, necessitating protein-specific experimental maps to fully understand biological regulation and design advanced therapeutics.

More like this