Analytical ultracentrifugation
High-speed runs over a few hours, scanning the moving boundaries; analysed by c(s) or related Lamm-equation fits to give a distribution of sedimentation coefficients, the amount of each species, and an estimate of mass and shape.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
Sedimentation velocity
High-speed runs over a few hours, scanning the moving boundaries; analysed by c(s) or related Lamm-equation fits to give a distribution of sedimentation coefficients, the amount of each species, and an estimate of mass and shape.
How it differs from the standard method
- 40,000–60,000 rpm for typical proteins; 12 mm double-sector cells with ≈ 400 µL.
- Load 0.3–1.0 OD for absorbance optics, or use interference optics for high or non-absorbing samples.
- Fit the meniscus, f/f0 and systematic noise; report rmsd and s20,w.
1Watching molecules sediment
An analytical ultracentrifuge spins a sample in a sector-shaped cell with quartz windows while optics record the concentration at every radius. At up to 60,000 rpm a protein feels a force a quarter of a million times gravity.
Sedimentation velocity (SV)
High speed. Each species forms a moving boundary; its rate gives the sedimentation coefficient s and its spreading gives diffusion. A c(s) analysis fits the whole set of scans with the Lamm equation to give a distribution of s-values.
Sedimentation equilibrium (SE)
Low speed, for a day or more, until sedimentation and diffusion balance. The concentration then rises exponentially towards the bottom, and the steepness of that exponential is the buoyant mass, independent of shape.
2From s to size and mass
The sedimentation coefficient, in svedbergs (1 S = 10⁻¹³ s), is s = M(1 − v̄ρ) / (N_A·f). For a given mass, an elongated or flexible molecule has more friction (a larger frictional ratio f/f0) and sediments more slowly.
3How it fails
- A rotor that is not at temperature: the sample sediments too fast and convects, smearing the early boundaries.
- A meniscus in the wrong place in the fit: every s-value shifts and the residuals show a pattern.
- Systematic noise left in absorbance data: spurious peaks in c(s). Fit time-invariant and radial-invariant noise.
- Too dilute or too concentrated for the optics: minor species vanish into the noise, or the plateau exceeds the linear range of absorbance (about 1.2–1.5 OD).
- Equilibrium analysed before scans superimpose: the mass is too low. Long columns take days to equilibrate.
- An unmeasured baseline or a v̄ that ignores glycans, lipids or detergent: masses that are wrong by a fixed factor or that drift with concentration.
Key terms
Practise these as flashcards →Velocity per unit centrifugal field, in svedbergs; s20,w is corrected to water at 20 °C so values compare between buffers.
Friction relative to a compact sphere of the same mass: about 1.2–1.3 for globular proteins, higher for elongated, flexible or glycosylated ones.
The distribution of sedimentation coefficients that best fits all velocity scans with the Lamm equation, with diffusion tied to s through f/f0 (Schuck, SEDFIT).
Volume increment per gram of solute, ≈ 0.73 mL/g for proteins and lower for carbohydrate; it sets the buoyancy factor 1 − v̄ρ.
Air–solution boundary at the top of the column, the origin from which sedimentation is measured; fitted in the analysis.
A fixed radial pattern in absorbance scans from windows and optics, removed algebraically during fitting.
M(1 − v̄ρ)ω²/RT, the slope of ln c against r²/2 at equilibrium.
When to use it
- Quantifying monomer, dimer and aggregate in a purified protein or a biologic without a column matrix that could filter aggregates or disrupt complexes.
- Measuring the stoichiometry and affinity of self-association or of heteromeric complexes in solution, including weak interactions.
- Determining an absolute mass, or the shape implied by s and f/f0, when SEC elution is misleading (elongated, glycosylated or detergent-solubilised proteins).
Compared with alternatives
Faster and needs less expertise; gives absolute mass per peak, but the column can remove aggregates and dissociate weak complexes.
A minute-scale aggregation check in a cuvette; detects but cannot quantify or resolve species as AUC does.
Gives shape (Rg, Dmax, envelopes) for a monodisperse sample; AUC tells you first whether it is monodisperse.
Preparative spins, including ultracentrifugation and gradients, separate and collect material; analytical runs measure it in flight and recover nothing.
Sub-techniques
Choosing one here switches the whole page, including the protocol and the lab, to it.
High-speed runs over a few hours, scanning the moving boundaries; analysed by c(s) or related Lamm-equation fits to give a distribution of sedimentation coefficients, the amount of each species, and an estimate of mass and shape.
- 40,000–60,000 rpm for typical proteins; 12 mm double-sector cells with ≈ 400 µL.
- Load 0.3–1.0 OD for absorbance optics, or use interference optics for high or non-absorbing samples.
- Fit the meniscus, f/f0 and systematic noise; report rmsd and s20,w.