Prot Bio Lecture 12 Structure Predictions and Protein Folding

Computational protein structure prediction

The use of algorithms and computational models to determine the three-dimensional structure of a protein from its amino acid sequence, often integrating physical principles, statistical potentials, and known structural databases

Protein conformation

The specific three-dimensional arrangement of atoms in a protein, determined by rotations around backbone and sidechain bonds and stabilized by intramolecular interactions

Protein folding

The process by which a linear polypeptide chain acquires its functional three-dimensional structure through a combination of local and long-range interactions

Protein motifs

Recurring structural patterns formed by combinations of secondary structures that often have functional or evolutionary significance

Protein domains

Independently folding, structurally distinct regions of a protein that often correspond to specific functional units

Protein folding dynamics

The study of time-dependent processes and intermediate states that occur as a protein transitions from unfolded to folded states

ROSETTA

A computational protein structure prediction program that assembles protein structures by combining small fragments from known protein structures and optimizing them using Monte Carlo sampling and energy scoring functions

Monte Carlo methods

Computational algorithms that use random sampling to explore conformational space and identify low-energy protein structures

ROBETTA server

A web-based platform that integrates multiple structure prediction methods, including ROSETTA, to generate automated protein structure predictions

LINUS

A protein folding algorithm that predicts structure using a hierarchical, a priori approach by folding short segments first and then combining them into larger structures based on favorable interactions

CASP (Critical Assessment of Structure Prediction)

A community-wide experiment in which researchers predict protein structures before experimental structures are released, allowing unbiased evaluation of prediction methods

CAFASP (Critical Assessment of Fully Automated Structure Prediction)

A variant of CASP focused on evaluating fully automated prediction pipelines without human intervention

CAPRI (Critical Assessment of Protein Interactions)

A blind prediction experiment focused on modeling protein-protein interactions and complexes

AlphaFold

A deep learning-based protein structure prediction system that uses neural networks trained on known structures to predict highly accurate 3D protein models from sequence

AlphaFold 1

The initial version that used deep learning to improve structure prediction but relied heavily on existing datasets

AlphaFold 2

An advanced version that uses attention-based neural networks and evolutionary information to predict structures with near-experimental accuracy

AlphaFold 3

A newer version capable of predicting protein complexes, including interactions with nucleic acids, ligands, and ions

Deep learning

A subset of machine learning involving neural networks with many layers that can identify complex patterns in large datasets

Evolutionary correlations

Patterns of co-evolution between amino acids in protein sequences that indicate spatial proximity in 3D structure

AlphaFold Protein Structure Database

A large database containing predicted protein structures covering most known protein sequences

Pattern (fold) recognition

Identifying structural similarities between proteins based on conserved folding patterns even when sequence similarity is low

Local physical refinement

Adjustment of predicted structures using energy minimization and physical constraints to improve accuracy

Protein Data Bank (PDB) dataset

A collection of experimentally determined protein structures used to train predictive models like AlphaFold

Alpha helix

A right-handed helical secondary structure stabilized by hydrogen bonds between backbone carbonyl oxygen and amide hydrogen atoms four residues apart

Beta sheet

A secondary structure formed by hydrogen bonding between extended polypeptide strands arranged in parallel or antiparallel orientations

Parallel beta sheet

A beta sheet in which adjacent strands run in the same N-to-C direction, resulting in angled hydrogen bonds

Antiparallel beta sheet

A beta sheet in which adjacent strands run in opposite directions, resulting in more linear and stable hydrogen bonds

Polyproline II (PII) helix

A left-handed helical structure with a three-residue repeat, commonly found in collagen and characterized by extended conformation

Turn or loop

A non-repetitive region of a protein that connects elements of secondary structure and often resides on the protein surface

Random coil

A flexible, disordered conformation lacking stable secondary structure

Protein structure visualization

Methods for representing protein structures, including ribbon diagrams, space-filling models, and backbone traces, to highlight different structural features

Phi (φ) angle

The dihedral angle around the N–Cα bond in the protein backbone that contributes to backbone conformation

Psi (ψ) angle

The dihedral angle around the Cα–C bond in the protein backbone that determines chain direction

Omega (ω) angle

The dihedral angle around the peptide bond, typically restricted to ~180° due to partial double-bond character

Sidechain conformations

The spatial orientations of amino acid side chains determined by rotation around χ (chi) angles

Rotamer libraries

Collections of preferred sidechain conformations derived from statistical analysis of known protein structures

Proline conformational restriction

The unique cyclic structure of proline limits rotation around its backbone, influencing protein folding and secondary structure formation

Peptide bond planarity

The peptide bond has partial double-bond character, restricting rotation and enforcing a planar geometry in the backbone

Backbone rotation constraints

Although many theoretical conformations exist, steric hindrance and chemical constraints limit allowable conformations to a small subset

Levinthal paradox

The observation that a protein cannot sample all possible conformations randomly due to astronomical time requirements, implying that folding follows specific pathways

Conformational space

The total set of possible structures a protein can adopt based on bond rotations

Hierarchical folding

A model in which local secondary structures form first and guide the formation of the overall tertiary structure

Ramachandran plot

A graph showing allowed and disallowed regions of phi and psi angles based on steric constraints, used to evaluate protein conformations

Allowed regions

Regions of phi/psi space where steric clashes are minimized and stable conformations exist

Favorable regions

Subsets of allowed regions corresponding to energetically preferred conformations

Steric hindrance

The physical restriction of molecular conformations due to atomic crowding and overlap

Disulfide bonds

Covalent bonds formed between cysteine residues that stabilize protein structure

Hydrogen bonds

Noncovalent interactions between electronegative atoms and hydrogen that stabilize secondary and tertiary structures

Hydrophobic effect

The tendency of nonpolar side chains to cluster in the protein interior to minimize disruption of hydrogen-bonded water networks

Van der Waals forces

Weak attractive forces between closely packed atoms that contribute to protein stability

Dense packing

The tight arrangement of atoms in a protein interior that maximizes stabilizing interactions

Supersecondary structures

Combinations of secondary structure elements that form recognizable patterns, such as beta-alpha-beta motifs

Structural motifs

Recurring arrangements of secondary and tertiary structures that are conserved across proteins

Protein fold

The overall three-dimensional arrangement of secondary structure elements in a protein

Protein domain

A compact, independently folding unit within a protein that often corresponds to a specific function

Modular protein structure

The concept that proteins are composed of multiple domains that can be recombined evolutionarily to create new functions

EGF domain

A protein domain found in epidermal growth factor involved in signaling

TPA (tissue plasminogen activator) domain

A domain involved in proteolysis and blood clot breakdown

Molten globule

A partially folded intermediate state characterized by native-like secondary structure but loosely packed tertiary structure

Folding funnel

A conceptual energy landscape in which unfolded proteins move toward a low-energy native state through multiple pathways

Framework model

A folding model where secondary structures form first, followed by assembly into tertiary structure

Hydrophobic collapse model

A folding model in which rapid burial of hydrophobic residues leads to formation of a compact intermediate before refinement of structure

Nucleation-condensation model

A folding model where key interactions form early and guide simultaneous development of secondary and tertiary structures

Oil drop model

A model describing protein structure as having a hydrophobic core surrounded by a more polar surface, though not all hydrophobic residues are buried

Charged residue exposure

Charged amino acids are typically located on protein surfaces due to favorable interactions with water

Initial collapse

A rapid early stage of folding where the polypeptide chain contracts into a compact form

Secondary structure formation

The formation of alpha helices and beta sheets early in folding

Tertiary structure formation

The establishment of long-range interactions that stabilize the full 3D structure

Native state

The fully folded, functional conformation of a protein

Folding speed limit

The theoretical fastest rate at which a protein can fold, approximately proportional to chain length (N/50 microseconds)

Folding intermediates

Transient structures formed during the folding process before reaching the native state

Folding timescale

Protein folding typically occurs within milliseconds to seconds depending on size and complexity