Introduction to Analytical Ultracentrifugation

This document relates to the analytical ultracentrifuge in the Xu lab at the Life Sciences Institute. The specifications and restrictions listed here are not necessarily indicative of machines located elsewhere.

For those who want to know, the instrument in question is a Beckman XLI-A analytical ultracentrifuge. This is a centrifuge capable of speeds in excess of 50,000 rpm, and contains an optics module capable of taking real-time readings of a sample while spinning. It uses a specific rotor with corresponding sample cells. There are two primary experiments for which this experiment can be used.

#1 - Velocity sedimentation - A sample is run at high speed, and the relative or absolute concentration at each radial point is measured with respect to time. Useful for partially or fully purified samples when little is known about the sample. The base output is a plot of absorbance (or fringe displacement, if using interference optics) vs. radius. Base output at multiple time points is combined to form one data set.

#2 - Equilibrium sedimentation - Samples (single or multiple) are run at low speeds for long periods of time until an equilibrium is established, and an absorbance reading across the cell is taken. Useful for nearly homogenous samples where molecular weight and/or molecule composition is known to determine association constants and, in some cases, more complex values (enthalpy, entropy). Output is absorbance vs. radius, and multiple sample concentrations and speeds are combined to make a single data set.

What does it do?

What does it not do?

- The AU is in no way preparative. Sample cells are small - velocity cells hold about 0.5 ml per chamber, while equilibrium cells hold about 0.15 ml per chamber (six chambers per cell). In the case of equilibrium cells, what goes in does not come out. You cannot reliably extract a sample from an equilibrium cell once it is closed. This is due to the design of the cell, and cannot be helped. Samples can be harvested from velocity cells fairly easily, but not in a fractionated way - there is no reliable way to get to the middle of the cell and draw out an aliquot, and no way to slow-drip the sample out. If you are looking to separate a sample in solution and extract it for subsequent assays, I suggest you try gel filtration chromatography.

- The AU typically will not be able to handle crude lysates. Equilibrium, in particular, would be a complete mess to try to decipher. Velocity runs may be able to deconvolute a large number of particles, but the more particles there are, the less chance you have of extracting data from the one you want. Typically, some sort of initial purification needs to take place before you run anything in the AU.

- The AU is not specifically designed to handle sucrose or cesium chloride gradients. While there's no problem with doing so, the software for analyzing data does not take this into account when calculating size or kinetic constants. When running experiments in this machine, Our lab always looks for sedimentation to the bottom of a cell, not to a midpoint in a gradient. If you think you have a pretty good handle on the experiment, you are more than welcome to use the machine for these purposes, but we won't be able to help you interpret the data.

Who do I talk to?

At the moment, the person most directly responsible for training, operation and maintenance of the AU is Anthony Ludlam in the Xu lab. Secondary operator in the Xu lab is Junyu Xiao (you can find contact information on the Xu lab web page). Please note that, due to the sensitive nature of the machine and the cost of replacement for damaged parts, we are trying to keep the number of "hands on" users to a minimum. Thus, we are enforcing a "one user per lab" policy - if you want to use the AU, you will be the only person from your lab using the AU until you vacate your position. Exceptions are made on a per-individual basis, based upon prior experience and frequency of use.

How do I analyze my data?

That's completely up to you. If you need help, you can talk to Anthony Ludlam, or if you want the data analyzed for you talk to Zhaohui Xu about setting up a collaboration. Typically it's not that hard to figure out how to analyze data, though. At the time of writing this document, our lab exclusively uses "Ultrascan", written by Borries Demeler at the University of Texas (http://www.ultrascan.uthscsa.edu/). This program is available free to academic users for windows, unix, and linux platforms. A Mac OSX version should be available shortly. From personal experience, the windows version tends to crash when handling too many data traces or doing extended monte carlo analysis, but that may be an artifact of the operating system. You can also contact Borries directly and have him analyze your data professionally - he does this work for free in exchange for publications rights (he'll provide figures, analysis, and method), or will charge $50 per cell on a fee-for-hire basis.

We have also heard good things about "WinNonLin" (http://spin6.mcb.uconn.edu/) as a fitting program for equilibrium data, and "sedfit" (http://www.analyticalultracentrifugation.com/default.htm) for velocity data. The former has a free 30-day trial and costs around $40 to register (status may have changed since we last checked). Our lab has no experience with the latter.

A copy of Ultrascan is located on the computer that runs the analytical ultracentrifuge, and you are welcome to use it so long as there is not a sample running at the time - when Ultrascan crashes, it will occasionally crash the entire system, which would disrupt a run in progress and introduce errors in the data.