Protein Folding Studies Using Fluorescence
Rapid-Scanning Emission Spectral Acquisition with an: Olis RSM 1000 plus Olis U.S.A. Stopped-Flow Optimized for Fluorescence.
The following data were acquired by Dr. Michael Schimerlik of Oregon State University and Dr. Richard DeSa using samples provided by Dr. Schimerlik. Work was done during Dr. Schimerlik's visit to Olis on October 17, 1995. In addition, Biophotonics International produced a six-page article in 1996 titled "Spectrofluorometry Helps Demystify Protein Folding."
T4 lysozyme refolding by combining pH jump and urea dilution.
(A) T4 lysozyme unfolded in urea
(B) pH 7 phosphate buffer
for final protein concentration of approximately 0.1 mg/ml.
Mixing A & B dilutes the urea and changes pH so that the protein folds.
An Olis RSM 1000 plus Olis U.S.A. stopped-flowfitted with a 450-watt xenon arc lamp as the exciting source, a single grating monochromator for selecting the excitation wavelength, and the middle-plane photomultiplier tubewas set for 0.3 seconds of data acquisition at one scan per millisecond.
The excitation ('sister') monochromator was set to 280 nm with 20 nm bandpass. The emission monochromator (the RSM or DeSa monochromator) was set to scan a 150 nm span with approximately 10 nm bandpass.
The experimentfrom collection of data to presentation of the rate constanttook 10 seconds.
150 microliters per shot of each reagent were used.
"DVM/Live Mode" within the Olis Software
Above are nine scans (2D data points) from the 500 collected during the course of the 0.3 second reaction (ISO were acquired as "pre-trigger" information); any nine scans could have been selected for display from among the 500. The "Scan Index" is marked with colored lines to illustrate where along the index of scans the displayed ones were extracted.
On the right is the kinetic trace at 369 nm, which was extracted from the 500 scans; the kinetics at any wavelength could have been extracted and the software supports multiple displayed traces simultaneously.
The kinetic trace shows "pre-trigger" or "pre-flow" data to the left of "time 0," where "time 0" is defined as the stopping of flow. We recommend that pre-trigger data be acquired, as it contains information about the flow process. If bubbles or other contaminants are present, they will be seen in the pre-trigger data.
One sees the rise in fluorescence as the unfolded protein was injected into the observation cell (the pre-time 0 data) and the subsequent quenching of fluorescence as the protein refolds (the post-time 0 data).
Result of 2D Fit to Kinetics at One Wavelength
The kinetic trace extracted from 369 nm is fitted to a single exponential. The upper plot is the residuals of the fit (i.e., the difference between the raw and fitted data). This kinetic trace is shown to be a single exponential, as confirmed by the near-perfect randomness of the noise in the residuals (plot above raw data graph).
The numeric information confirms the graphic. The error in the rate and amplitude are good, given the high level of noise in the raw data. And, very importantly, the Durbin-Watson ratio ("DW") of 2.1/1.65 is amazingly high; for a good fit, this statistical ratio value should be "1" or better, affirming statistically that the noise is random.
The returned rate is 21±3. The relatively high uncertainty is because of the noise level. But recall! These data were collected in less than 0.3 seconds and were collected as a function of wavelength! Each datum on the kinetic trace came from a one-millisecond emission scan! (See also note about five-fold improvements t signal-to-nose on RSM new procedures.)
The figure below shows how much better the answers are when all of the data are used to calculate the rate constants rather than just a single wavelength's. As we explain in our Robust Global Fitting write-ups , RGF includes a step called "factor analysis," which identifies and isolates noise contributions. Thus, while we provide 2D fits in the RSM software, we recommend one use this facility only rarely with good cause. Using all wavelengths (rather than isolating and fitting just one wavelength) and using RGF (rather than 2D fitting) is to take full advantage of the capabilities of an Olis Spectrofluorimeter.
The Answers Returned by Robust Global Fitting
The right graph, containing spectral displays, shows the emission spectrum of the unfolded protein, i.e., the product of the reaction, and the starting spectrum, i.e., the folded protein. We see that both the intensity of the fluorescence (amplitude) and the peak wavelength changed as the protein refolded.
The left graph, containing kinetics, shows the appearance of one species with the simultaneous disappearance of another species. Above these traces is a third, which shows the overall kinetic change. And above these is the plot of the residuals, almost perfectly distributed about zero.
Below the plots are the numeric answers, including the overall standard deviation and the reported rate constant of 18.5±0.7. Notice the marked improvement to the returned rate constant and error.
With one stopped-flow shot, spectral acquisition for 0.5 seconds, and less than 5 seconds of pressing keys and computer calculation time, you have the answer to a protein folding experiment! There is one fluorimeter in the world capable of this time-resolution and calculation powerthe Olis RSM 1000F