How Protein Folding Works
🦴 Primary Structure
A protein is a polypeptide chain — amino acids linked by peptide bonds (–CO–NH–). The sequence of residues (primary structure) entirely determines the final folded shape, as shown by Anfinsen's experiment (Nobel 1972).
🌀 Secondary Structure
Local hydrogen bonding patterns create regular motifs: α-helices (phi ≈ −57°, psi ≈ −47°) coил one backbone H-bond per 3.6 residues; β-sheets (phi ≈ −120°, psi ≈ +130°) form extended parallel or anti-parallel strands.
⚡ Energy Minimisation
Folding is driven by minimising Gibbs free energy: ΔG = ΔH − TΔS. The hydrophobic effect (burial of nonpolar sidechains) is the dominant force. Van der Waals, electrostatics, and disulfide bonds stabilise tertiary structure.
🔬 Folding Funnel
The energy landscape is a rugged funnel with the native state at the global minimum. Chaperone proteins (Hsp70, GroEL/ES) prevent off-pathway aggregation by binding hydrophobic patches on partially folded chains.
Protein Structure Comparison
| Structure Type | φ angle | ψ angle | H-bonds per residue | Rise per residue | Residues per turn | Example |
|---|---|---|---|---|---|---|
| α-Helix (right-handed) | −57° | −47° | 1.0 (i to i+4) | 1.5 Å | 3.6 | Haemoglobin subunits |
| 3₁₀-Helix | −49° | −26° | 1.0 (i to i+3) | 2.0 Å | 3.0 | C-terminal caps |
| β-Sheet (antiparallel) | −139° | +135° | Interstrand | 3.4 Å | — | Immunoglobulin domains |
| β-Sheet (parallel) | −119° | +113° | Interstrand | 3.2 Å | — | TIM barrel enzymes |
| Polyproline II | −75° | +145° | None | 3.1 Å | 3.0 | Collagen triple helix |
| β-Turn | Variable | Variable | 1 (i to i+3) | — | — | Hairpin loops |
| Random Coil | Variable | Variable | None regular | — | — | Intrinsically disordered |
| Amyloid Fibril (misfolded) | −118° | +130° | Cross-β interstrand | 4.7 Å | — | Aβ plaques (Alzheimer's) |
Frequently Asked Questions
What drives protein folding?
Protein folding is driven by the hydrophobic effect (burial of nonpolar residues), hydrogen bonding, van der Waals forces, and electrostatic interactions. The net result is minimisation of Gibbs free energy: ΔG = ΔH − TΔS. At physiological temperature (~310 K) the entropic cost of ordering is more than offset by enthalpic gains from contact formation.
What is a Ramachandran plot?
A Ramachandran plot shows the phi (φ) and psi (ψ) backbone dihedral angles of amino acid residues. Sterically allowed regions cluster around α-helix (φ ≈ −57°, ψ ≈ −47°) and β-sheet (φ ≈ −120°, ψ ≈ +130°) conformations. Glycine occupies the widest range (no side chain); proline is restricted to a narrow region. High-quality crystal structures have >98% residues in allowed regions.
How does protein misfolding cause disease?
Misfolded proteins can aggregate into amyloid fibrils or prion structures. In Alzheimer's disease, misfolded amyloid-β and tau aggregate into plaques and neurofibrillary tangles. In prion diseases, normal PrPC converts to infectious PrPSc through contact-induced misfolding. Type 2 diabetes involves islet amyloid polypeptide (IAPP) fibril deposition in pancreatic β-cells.
What is the Levinthal paradox?
Cyrus Levinthal (1969) noted that if a 100-residue protein sampled all possible conformations randomly (assuming 3 states per bond), it would take ~1047 years to fold — yet proteins fold in microseconds to seconds. The paradox is resolved by energy funnel landscapes: directed gradients guide chains efficiently toward the native state without exhaustive conformational search.
🧬 Explore Molecular Biology
Deepen your understanding with our full-length articles on protein structure and computational biology.
Protein Structure & Function → Bioinformatics & Computational Biology →