What is protein folding?
A protein is a chain of amino acids. Within microseconds to seconds, most chains spontaneously fold into a specific 3D shape – the shape that lets the protein work as an enzyme, antibody, motor or building block.
The sequence of amino acids (the primary structure) contains the information for the final fold. This simulator reduces each residue to a single bead so the physics of folding is easy to see.
Levels of structure
- Primary – the amino-acid sequence, linked by peptide bonds.
- Secondary – local patterns: α-helices and β-strands/sheets, held by backbone hydrogen bonds.
- Tertiary – the full 3D arrangement of one chain, stabilised by a buried hydrophobic core.
- Quaternary – several chains assembling into a complex (e.g. haemoglobin has four).
Forces that drive folding
- Hydrophobic effect – non-polar side chains are expelled from water and cluster in the core. It is usually the dominant driving force, and is shown by the orange beads.
- Hydrogen bonds – N–H···O=C bonds between backbone groups build helices and sheets.
- van der Waals packing – tight, complementary packing in the core releases small but numerous attractions.
- Electrostatics – positive and negative residues form salt bridges, screened by salt in the solvent.
- Chain entropy – folding costs conformational freedom; the energy gain must outweigh this penalty. That is why high temperature unfolds proteins.
Energy landscapes & Levinthal's paradox
A 100-residue chain has an astronomical number of conformations (~1047 with just 3 states per bond). Searching them randomly would take longer than the age of the universe – yet proteins fold in milliseconds. This is Levinthal's paradox.
The resolution is that folding is biased: the energy landscape is funnel-shaped. Every contact that forms lowers the energy and guides the chain downhill. Rough funnels have traps (misfolded states, molten globules) where the chain can get stuck – try a low specificity or temperature.
Misfolding & disease
When folding goes wrong, proteins can aggregate into amyloid fibrils, linked to Alzheimer's, Parkinson's and prion diseases. Cells use chaperones (such as GroEL and Hsp70) to give misfolded proteins another chance, and the proteasome to destroy those that never fold.
About this model
Each bead is one residue (the Cα atom), bonds are rigid (~3.8 Å). Motion follows Brownian (Langevin) dynamics: random thermal kicks plus forces from the energy function:
E = Eexcluded + Ehydrophobic + Enative + Eelectrostatic + Elocal
Native terms form a Gō-like model: residues in contact in the target structure attract each other, representing the interactions encoded by the real sequence. Specificity blends between this designed attraction (1) and generic hydrophobic collapse (0). Energies are in reduced units; one unit ε ≈ 0.6 kcal/mol. This is a teaching model – real proteins have side chains, explicit water and hydrogen bonds, and are far more complex.
Try these experiments
- Pick the Three-Helix Bundle, press Unfold and watch the core form.
- Raise Temperature above ~0.9: the fold melts (denaturation). Cool it to refold.
- Drop Native specificity to 0: the chain collapses to a compact but disordered molten globule.
- Choose β-Sheet at low temperature – it can get kinetically trapped. Try Heat & refold (annealing).
- Click a bead to inspect its neighbours and interactions.