TL;DR
This paper explores why some drugs bind to many different protein targets (promiscuity) while others are highly specific. By analyzing the 3D shapes of protein kinase binding sites, the authors identify key structural features that dictate binding affinity to the promiscuous inhibitor staurosporine, offering insights into designing more selective drugs.
What Problem Does This Paper Address?
Protein kinases are crucial for cell signaling, and their dysregulation is implicated in diseases like cancer. While many drugs target these kinases, achieving selectivity is a major challenge. Promiscuous inhibitors, like staurosporine, bind to a wide range of kinases, making them useful research tools but problematic as therapeutics due to off-target effects. Understanding the structural basis for this promiscuity is key to designing drugs that hit only the intended target. This paper aims to uncover the specific geometric and chemical features of kinase binding sites that contribute to staurosporine's broad binding profile.
Methods
The researchers employed a multi-pronged computational approach:
- Shape Correlation (Mantel Test): They used the Mantel test, a statistical method, to compare matrices of distances between key atoms in the binding sites of various protein kinases. This allowed them to group kinases based on the similarity of their binding site "shapes" (quasi-shapes), independent of their amino acid sequence.
- Grid-Based Analysis (MAHORI): They developed a web-based tool called MAHORI (Mapping Analogous Hetero-atoms onto Residue Interactions) to visualize and analyze conserved atomic environments around ligands. This involved creating a 3D grid around the bound ligand and identifying frequently occurring atom types and residue types at specific locations.
- Quantitative Structure-Activity Relationship (QSAR): They used multiple linear regression to identify specific distances between binding site atoms that correlated with staurosporine's binding affinity (Kd,STU). This helped pinpoint geometric features that influence how tightly a kinase binds the inhibitor.
Datasets
The study utilized data from the Protein Data Bank (PDB) for X-ray crystal structures of protein kinases. Specifically, they analyzed:
- Structures of protein kinases co-crystallized with staurosporine.
- Structures of protein kinases co-crystallized with ATP or adenine-containing compounds.
- A dataset of 35 non-redundant protein kinases and 10 inhibitors from a previous study (Fabian et al., 2005), which provided dissociation constants (Kd) for staurosporine and other ligands.
Key Insights and Ideas
- Shape Trumps Sequence: The study highlights that the 3D shape of a kinase's binding pocket can be a stronger predictor of binding affinity for certain ligands (like staurosporine) than the amino acid sequence alone. This explains why kinases from different families can sometimes bind the same promiscuous inhibitor.
- The Gatekeeper and DFG Loop are Crucial: The size of the "gatekeeper" residue and the proximity between the first glycine of the GXGXXG motif and the aspartate of the DFG loop are identified as key structural determinants for tight staurosporine binding. These elements appear to work in concert to "lock" the inhibitor in place.
- Hinge Region is the Constant: The main chain of the hinge region, which forms interactions with both staurosporine and ATP, is the most conserved part of the binding site across different kinases. This conserved interaction is a primary reason for staurosporine's promiscuity.
- Induced Fit for Glycine: The glycine-rich loop, often considered flexible, appears to undergo an induced fit or conformational selection upon staurosporine binding, becoming more fixed in position. This suggests that the ligand actively influences the protein's conformation.
- Methylamine Interactions Matter: The number of hydrogen bonds and ionic interactions made by the kinase around the methylamine group of staurosporine directly correlates with tighter binding. This points to specific functional groups on the inhibitor as key interaction points.
Key Results
- A shape-based dendrogram successfully clustered kinases with similar binding profiles to staurosporine, often separating them from clusters based on sequence similarity.
- Multiple linear regression identified specific distance descriptors that significantly correlated with staurosporine binding affinity, with R² values around 0.7 for test sets.
- The analysis revealed that the main chain of the hinge region is the most conserved structural element in the binding site for both staurosporine and ATP.
- Table 2 shows that kinases with more hydrogen bonds or ionic interactions to the N4 of staurosporine generally exhibit lower (tighter) binding affinities. For example, CDK2 and PKA, with 2 such interactions, have affinities below 51 nm, while STK16, with none, has an affinity of 200 nm.
Limitations and Future Work
- Apo Structures Needed: The authors acknowledge that further evidence from apo (unbound) structures is needed to definitively confirm the induced-fit mechanism observed for the glycine-rich loop.
- Predictive Power: While the QSAR models showed predictive power, the authors suggest that a larger dataset would be beneficial to further validate and refine these equations.
- Drug Design Implications: The study provides a framework for understanding inhibitor selectivity and promiscuity. Future work could involve applying this approach to design more selective inhibitors by targeting specific structural features identified in this study, or by modifying staurosporine to enhance interactions with key residues. The MAHORI tool is also presented as a resource for further exploration.
Summary updated