Bioinformatics: lecture 7+8

Residue Solvent Accessibility

• Solvent accessible surface area (ASA, SASA, or SAS) is measured in Ų.
• It quantifies how much of a residue in a protein structure is accessible to the solvent.
• It's calculated by rolling a sphere with a specific radius over the protein's surface and summing up the accessible area for each residue.
• Solvent excluded surface (SES) is also called molecular surface or Connolly surface area.
• It's often displayed in "surface" visualizations.
• Relative accessible surface area (rASA) is the ratio of the actual accessible surface area (ASA) of a residue to the maximum possible accessible surface area (ASAMAX) of that residue.
• It's used to compare the accessibility of different amino acids (e.g., long extended vs. spherical amino acids).
• A simplified two-state description of residue accessibility is "buried" vs. "exposed".
• A threshold (usually 15-25%) is used to differentiate between surface and buried residues.
• rASA values below the threshold indicate a buried residue, while values equal to or above the threshold indicate an exposed residue.

Protein Solubility

• Protein solubility is defined as the concentration of protein in a saturated solution that's in equilibrium with the solid phase.
• Several factors influence protein solubility for proteins expressed in a lab setting.
• Hydrophilic/hydrophobic balance of solvent-exposed residues is key.
• Aggregation-prone regions (APRs) play a crucial role, primarily containing hydrophobic residues prone to forming beta-structures.
• Protein expressibility in cells is also important.

Molecular Interactions

• Intramolecular interactions occur within the same protein structure.
• Intermolecular interactions happen between different proteins in assemblies.
• Understanding these interactions is crucial for understanding protein function and stability, as well as the stability and function of protein complexes.

Types of Interactions

• Charge-charge (ionic) interactions are present in charged residues, exemplified by salt bridges.
• Hydrogen bonds (H-bonds) involve a hydrogen atom shared between a donor and an acceptor atom.
• Aromatic (π-π) interactions are attractive interactions between aromatic rings.
• Van der Waals (vdW) interactions occur between any two atoms, being more important for non-polar residues.
• Hydrophobic interactions have an entropic origin and are important for non-polar/hydrophobic residues.
• Disulfide bonds (cysteine bridges) are formed by the oxidation of two cysteine residues.
• Cation-π interactions are electrostatic interactions between a positively charged residue (Lys or Arg) and an aromatic residue (Phe, Trp, or Tyr).
• Polar Interactions:
  • Arginine interactions, where positively charged Arg interacts with aromatic rings.
  • Arginine-arginine stacking where two Arg form parallel "aromatic" stacking.

Criteria for Recognizing Molecular Interactions

• Atom types/functional groups.
• Geometric rules (distances, angles).
• Energetics (physicochemical rules).
• Contact surface area between atoms.
• SASA (Solvent Accessible Surface Area).

Functional Sites

• Functional sites on proteins are regions providing complementarity for bound molecules (ligands).
• Binding sites are essential for molecular recognition, while active/catalytic sites promote chemical catalysis.
• Binding involves non-covalent interactions between protein and ligand.
• Bound molecules are small molecules or macromolecules and often display high specificity.

Binding Sites for Small Molecules

• Binding sites often form internal cavities, surface pockets, or clefts (concave regions).
• Microenvironments within binding sites differ from the bulk solvent, impacting binding.
• These sites are often highly conserved through evolution.
• Binding sites typically exhibit low desolvation energy and have characteristic physicochemical properties.
• Different methods for identifying binding sites include; Evolutionary conservation, physical detection of “pockets,” Geometry based methods, Energy based methods, Knowledge-based, Machine learning-based, Template-based, Microenvironment-based.

Binding Sites for Macromolecules

• Binding sites are typically protruding loops, large surface clefts, or flat sites; binding for macromolecules is flatter than small molecules.
• Recognition of macromolecules involves interactions over large, continuous surface areas or several discrete binding regions.
• Identifying these sites is often challenging due to their larger size and complexity.
• Binding sites are generally highly conserved in evolution.
• Low desolvation energy.
• Specific Physicochemical properties.
• DNA binding sites have characteristic motifs and positive charged electrostatic patches.

Approaches to Identify Binding Sites

• Evolutionary Conservation.
• Physical Detection of "pockets."
• Geometry-based methods.
• Energy-based methods.
• Knowledge-based methods.
• Template-based methods.
• Microenvironment-based methods.
• Meta-servers (combine multiple methods).

Knowledge-based Methods

• Combine multiple interface features (conservation, residue propensity, physicochemical properties, structural properties).
• Use known binding sites for parameterization or training empirical scoring functions.

Transport Pathways

• Transport pathways mediate the transport of ions and small molecules within proteins.
• Channels/pores facilitate transport across membranes.
• Tunnels facilitate ligand exchange between buried active/binding sites and the bulk solvent.
• Intramolecular tunnels carry reaction intermediates.
• The permeability to different substances depends on size, shape, amino acid composition, and dynamics.
• Bottleneck regions are critical for selectivity.
• Methods for predicting transport pathways include the identification of overall voids, tunnels, and channels.

Identification of Overall Voids

• Methods accurately represent various voids (channels, tunnels, surface clefts, pockets, and internal cavities.)
• Tools like HOLLOW and 3V help analyze protein surfaces.

Identification of Tunnels

• Methods calculate tunnels connecting occluded cavities to the surrounding bulk solvent.
• Voronoi diagrams analyze void spaces.
• Dijkstra's algorithm identifies optimal pathways.
• Probe size determines the minimum radius threshold.
• Tunnel geometry approximated by spheres.

Identification of Channels

• Methods calculate channels (or pores) that penetrate throughout the protein.
• They aren't suitable for identifying tunnels from occluded cavities.
• Usually consider only one channel per structure.
• Information about the channel's approximate position and direction might be needed.

Protein-Ligand Complexes

• Protein-ligand complexes are crucial for various biological processes.

Molecular Recognition

• Molecular recognition is the specific interactions between two or more molecules through non-covalent interactions; key for various biological roles (e.g., catalysis, signaling).
• Models are used to explain molecular recognition phenomena.

Lock-and-Key Model

• Proposed by E. Fisher in 1894, it posits a rigid receptor and ligand with complete complementarity for efficient binding.
• The model may not accurately portray allosteric effects or catalytic mechanisms.

Induced-Fit Model

• Proposed by D.E. Koshland in 1956, this model highlights that a partial complementarity may suffice for interactions.
• Both the ligand and receptor can undergo conformational adjustments.
• The conformation of the bound receptor often isn't found in the unbound state.

Selected-Fit Model

• This model, introduced by B.F. Straub in 1964, considers flexible receptor and ligands (ensembles).
• Complex formation is similar to a lock-and-key interaction when complementary configurations occur.
• The bound receptor's conformation can also exist in its free state.

Keyhole-Lock-Key Model

• Proposed by Z. Prokop in 2012, this model applies when the receptor has a buried active site and tunnels. It highlights that complementarity with the ligand is needed for both the active site and the tunnel and explains the increased selectivity of the tunnel.

Biocatalysis

• Enzymes accelerate chemical reactions by lowering the activation barrier.
• Provides environments to stabilize transition states thereby enhancing reaction rates.

Structures of Complexes

• Databases such as PDBbind, BindingDB, and ChEMBL provide experimentally determined complexes.

Protein Druggability

• Druggability is the likelihood of a protein being targeted by a drug-like molecule to elicit a therapeutic effect.
• Crucial for predicting if a biological target is a promising drug candidate).
• Lipinski's Rule of 5 is a guideline for orally active drugs and commonly used to assess drug-likeness.

Prediction of Protein Druggability

• Prediction methods use similarity to known targets, sequence/structural characteristics of binding domains, and data from databases of known targets.
• Predictive tools like PockDrug Server and DoGSiteScorer speed up the target identification phase of drug discovery.

Small Molecules

• Methods for representing ligands, databases like the Cambridge Structural Database (CSD), PubChem, and ZINC are important.
• Common methods used for representing small molecule structure include 1D (empirical formula); 2D (chemical structure diagrams), and 3D (atomic coordinates).

Molecular Docking

• Molecular docking is a crucial technique when experimental data for ligand binding is unavailable or for screening a large number of molecules.
• It has several stages and involves receptor representation, ligand representation, search for binding modes, and scoring to rank them.

Receptor Representation

• Receptor is represented by only the relevant binding site.
• Descriptor representation (obtained from geometry and interaction abilities).
• Grid representation (to cover the entire search region with relevant spatial information from the probe).
• Handling receptor flexibility, employing rigid/soft docking, identifying explicitly flexible side chains, or applying ensembles from experimental or simulated structures.

Ligand Representation

• Ligands are represented by all (or some selected) atoms.
• Non-polar hydrogens can be combined with their parent atoms
• Only single bond rotations are considered during the docking of ligands.
• Libraries of pre-generated conformations can be docked.
• Ligand flexibility is important. Fragments may be used for direct sampling of conformational space.

Molecular Docking—Search

• Various search algorithms are available.
• Geometry-based, combinatorial methods; assume that binding affinity is proportional to complementarity between ligand and receptor (which is not always the case, especially for polar ligands).
• Energy-driven and stochastic algorithms try to directly locate the global minimum of binding free energy.

Molecular Docking—Scoring

• Scoring functions evaluate all binding modes.
• Essential for assigning relative rankings to these modes.
• Different scoring function categories include empirical, knowledge-based, force field-based, and machine learning-based.

Evaluation of Complexes

• Intermolecular interactions (e.g., hydrogen bonds, hydrophobic interactions, aromatic interactions, ionic interactions).
• Binding energies.

Transport of small molecules.

• Describe trajectories of ligands through tunnels, based on geometry or force fields.
• Geometry-based approaches are fast but have low accuracy.
• Examples of geometry-based software include CaverDock, MoMA-LigPath, and SLITHER.
• Force field-based approaches use MD simulations, which are computationally expensive but offer higher accuracy.

Test your knowledge on protein structures, residue accessibility, and their interactions. This quiz covers topics like protein solubility, binding sites, and computational tools used in bioinformatics. Dive in to see how well you understand the intricacies of protein science!