Sedimentation velocity is an analytical ultracentrifugation (AUC) method that measures the rate at which molecules move in response to centrifugal force generated in a centrifuge. This sedimentation rate provides information about both the molecular mass and the shape of molecules. In some cases this technique can also measure diffusion coefficients and molecular mass.
In the biotechnology industry sedimentation velocity is used much more frequently than sedimentation equilibrium and thus when biotech scientists say "AUC" or "analytical ultracentrifugation" they typically really mean "sedimentation velocity".
Sedimentation velocity is particularly valuable for:
In the sedimentation velocity method a sample is spun at very high speed (usually 40-60 K rpm) in an analytical ultracentrifuge. The high centrifugal force rapidly depletes all the protein from the region nearest the center of the rotor (the meniscus region at the air/solution interface), forming a boundary which moves toward the outside of the rotor with time (see example below), until finally all the protein forms a pellet at the outside of the cell. The concentration distribution across the cell at various times during the experiment is measured while the sample is spinning, using either absorbance or refractive index detection in our Beckman ProteomeLab XL-I (picture).
A major advantage of this method over sedimentation equilibrium is that experiments usually require only 4-6 hours, as opposed to the several days typical of sedimentation equilibrium. Thus sedimentation velocity can be used with samples that are too labile for sedimentation equilibrium. The major drawback relative to sedimentation equilibrium applies to interacting systems (proteins that reversibly self-associate or protein-protein complexes), where the non-equilibrium nature of the measurement can lead to significant changes in species distributions over the course of an experiment. Further, for interacting systems it is generally more difficult and less accurate to derive binding constants (Kd's) from sedimentation velocity data.
An important strength of sedimentation velocity is its ability to study samples over a fairly wide range of pH and ionic strength conditions (and often directly in formulation buffers), and at temperatures from 4 to 40 °C. The amount of protein required depends on the application, but each sample is usually ~0.45 mL at typical protein concentrations of 0.1-1 mg/mL (45-450 micrograms total). Protein concentration can range as low as ~10 micrograms/mL or as high as ~40 mg/mL in some cases (but generally the concentration should be 2 mg/mL or below). Up to 3 samples can be run at one time.
To better illustrate this technique and some of its uses, some data for a monoclonal antibody will be presented here and the interpretation of that data briefly discussed. Even more example applications for antibodies can be found among the published articles, talks, and posters marked with the symbol on our Further Reading page.
The graph below shows scans across the centrifuge cell, recording the absorbance at 280 nm versus position within the cell. These scans were taken starting at 13 minutes after initiating a run at 45,000 rpm (the black data set in the graph), and then every ~12 minutes thereafter (blue, green, cyan, etc.). The sharp vertical spike at 6.02 cm indicates the position of the air-solution meniscus. In the first data set the sedimentation of the antibody has already depleted its concentration in the region near the meniscus and formed a sedimentation boundary. At later times in the run the depleted region expands and the boundary moves away from the center of the rotor, until by the time of the last data set the concentration of antibody has dropped to essentially zero throughout the upper half of the cell.
The rate at which the sedimentation boundary moves is a measure of the sedimentation coefficient of the protein. The sedimentation coefficient depends on the molecular weight (larger proteins sediment faster) and also on molecular shape. Unfolded proteins or one with highly elongated shapes will experience more hydrodynamic friction, and thus will have smaller sedimentation coefficients than a folded, globular protein of the same molecular weight.
The minimum width of the sedimentation boundary is related to the diffusion coefficient of the molecule; the presence of multiple species with similar sedimentation coefficients will cause the boundary to be broader than expected on the basis of diffusion alone. In this case the majority of the boundary is reasonably narrow, but the slow rise of the data on the right side of the boundary suggests the presence of some faster moving species.
When viewed as in the graph above the data is difficult to interpret. What we often want to know is how much material is sedimenting at various sedimentation coefficients. By taking many scans close together in time (the graph above shows only a small number of the scans), subtracting them in pairs, and doing some mathematical manipulation these data can be transformed into the sedimentation coefficient distribution, g(s*), which is shown below. This distribution resembles a chromatogram, and in many ways is similar to a size-exclusion chromatogram except the peaks come in the opposite order. Like a chromatogram, the area under each peak gives the total amount of that species.
For this antibody sample we see only one distinct peak, centered at a sedimentation coefficient of ~6.5 S, which corresponds to the native antibody 'monomer' (really a covalent heterotetramer of 2 light and 2 heavy chains). A sedimentation coefficient of 6.5 S is actually rather low for a 150 kDa species, which is consistent with high hydrodynamic friction from its highly asymmetric, non-globular 'Y' shape.
Although we see only a single peak in the g(s*) distribution, a more detailed analysis quickly reveals that this sample is not homogeneous. The red curve is a fit of these data as a single species. This fit clearly fails to account for the data over the region from 8-12 S, indicating the presence of some dimer and possibly also some trimer. The fact that the distribution has not returned fully to zero at 12 S also suggests that there may also be small amounts of even faster sedimenting species present.
From the width of the main peak we can also calculate the apparent diffusion coefficient of the monomer. In turn, from the ratio of sedimentation coefficient to diffusion coefficient we can calculate a mass of 151 kDa for this species, which matches the expected value well within 3-5% error expected for masses determined in this fashion.
Lastly, we can apply a more complex analysis algorithm to further enhance the resolution of the species distribution (essentially sacrificing information about diffusion coefficients and mass in order to get higher resolution), as shown in the figure below.
In this figure the full distribution is shown in the main graph, while the inset magnifies the vertical scale by 10X in order to better show the minor components. With the enhanced resolution we now see a fully baseline-resolved dimer peak at ~9.4 S (4.7% of the total protein), and small peaks at ~13.7 S and 17.5 S (1.8% and 1.1%, probably trimer and tetramer). In addition there is 0.7% of a low mass contaminant at 2 S (possibly free light chain).
Thus from this one velocity experiment we have been able to quantify the amount and mass of the main component, the content and sedimentation coefficients of 3 aggregates and a low mass contaminant, and to obtain information about the conformation of the main component.
To further illustrate the excellent resolution and sensitivity that can be obtained, the next figure shows results for a highly stressed sample of another monoclonal antibody. We see well-resolved peaks for dimer, trimer, ... out to heptamer, as well as 2 peaks for fragments. (The hexamer and heptamer peaks are too small to see on this scale.) This resolution is much better than can be achieved by size-exclusion chromatography, and a much broader range of sizes can be covered in one experiment via sedimentation velocity (this analysis could easily have been extended out to masses ~10-fold higher).
The figure below shows a sedimentation coefficient distribution for a sample of adenovirus (~150 MDa, used to deliver vectors for gene therapy). This illustrates the broad range of sizes and molecule types that can be studied using this technique. The inset shows a 20-fold expanded scale to allow the many minor peaks to be seen. This material turns out to be quite heterogeneous, having only 57% of the material as the main peak (virus monomer), plus many rapidly-sedimenting viral aggregates and some slowly-sedimenting species that may be fragments or empty capsids.
See also our AUC services page
All text, images, and downloadable documents are © copyright 2000-2013, Alliance Protein Laboratories Inc.