Intermediatemolecular biology
The Protein Folding Problem
Overview
A protein's function depends on its three-dimensional structure, yet the amino acid sequence alone encodes it. This lesson covers Anfinsen's experiments, the folding funnel model, and how AI cracked structure prediction.
Learning objectives
- State Anfinsen's dogma
- Explain the folding funnel model
- Describe what AlphaFold achieved
- Relate misfolding to disease
Anfinsen's dogma
Christian Anfinsen showed denatured ribonuclease refolds spontaneously: the amino acid sequence determines the native structure under physiological conditions (thermodynamic hypothesis, Nobel 1972).
Sequence → structure: the native fold is the thermodynamically most stable state.
Levinthal's paradox
A 100-residue protein has astronomical possible conformations — random search would take longer than the universe's age, yet proteins fold in milliseconds. Resolution: folding follows biased pathways down an energy funnel, not random search.
Funnel model resolves the paradox: local interactions progressively restrict the search space.
The folding funnel
The energy landscape is funnel-shaped: many unfolded states at the top, one native state at the bottom. Molten globules and intermediates are waystations; chaperones smooth the descent.
Folding is not a single pathway — it's a statistical descent with many parallel routes.
AlphaFold's breakthrough
DeepMind's AlphaFold2 (2020) predicts structures from sequence with experimental accuracy for most proteins, using evolutionary couplings learned from sequence alignments — effectively learning the folding code.
AlphaFold didn't simulate folding physics — it learned sequence→structure patterns from evolution.
When folding fails
Misfolding causes cystic fibrosis (ΔF508 CFTR), Alzheimer's (amyloid), and prion diseases — the subject of upcoming lessons.
See prion-misfolding-deep and amyloid-diseases for the dark side of folding.
Key points
- Sequence determines structure
- Folding funnel resolves Levinthal's paradox
- AlphaFold predicts from evolution
- Folding is fast and parallel
- Misfolding causes disease
Common misconceptions
✓ Truth: Levinthal's paradox shows that's impossible — folding follows energy-biased pathways.
✓ Truth: It predicts the final structure from evolutionary patterns, not by simulating folding dynamics.
Practice questions
Q1. Anfinsen's ribonuclease experiment demonstrated:
- Spontaneous refolding from sequence alone
- Chaperone requirement
- ATP-dependent folding
- Prion formation
Denatured ribonuclease refolded without help, proving sequence encodes structure.
Q2. Levinthal's paradox concerns:
- The impossibility of random conformational search
- Protein synthesis speed
- Enzyme kinetics
- DNA replication fidelity
The paradox: exhaustive search of conformations would outlast the universe, yet folding is fast.
Q3. The folding funnel's bottom represents:
- The native state
- The denatured state
- A molten globule
- An aggregate
The funnel narrows to the single lowest-energy native conformation.
Click a question to reveal the answer and explanation.
Frequently asked questions
What is Anfinsen's dogma?
The principle that a protein's amino acid sequence determines its native three-dimensional structure — demonstrated by reversible denaturation and refolding of ribonuclease.
How does the folding funnel resolve Levinthal's paradox?
The paradox notes random conformational search would take eons. The funnel model shows energy gradients progressively narrow the search — local contacts form early and guide the chain downhill, so folding takes milliseconds.
How does AlphaFold predict protein structures?
It uses deep learning on multiple sequence alignments: co-evolving residue pairs (which must be close in 3D) constrain the predicted structure, achieving near-experimental accuracy without simulating physics.