Dr. Chun-Biu Li, Associate Professor, Dr. Tamiki Komastuzaki, Professor and their colleagues in the Laboratory of Molecule & Life Nonlinear Sciences, Research Institute for Electronic Science, Hokkaido University have successfully indicated that some of the most commonly used data analysis methodologies have a great chance to cause misinterpreting results in the analysis of single enzymatic experimental time series.
The mechanisms by which enzymes achieve extraordinary reaction rate acceleration and specificity have long been one of most important key to understand the "algorithm" of living systems. Recent advances in single molecule (SM) experiments allow the real time observation of an individual enzyme during the time course of the catalytic reaction. It has been claimed (e.g. Nat. Chem. Bio., 2, 87 (2006)) that the single enzymatic turnover rates demonstrate multi-timescale fluctuations which may originate from the existence of multiple catalytic conformations with slow transitions among them.
Using a single molecule fluorescence approach, the time series of catalytic events of an enzymatic reaction can be monitored yielding a sequence of fluorescent "on"- and "off"-states. An accurate on/off assignment is complicated by the intrinsic and extrinsic noise in every SM fluorescence experiment (see fig.1). Using simulated data, the performance of the most widely employed binning and thresholding approach was systematically compared to change point analysis. It is shown that the underlying on- and off-histograms as well as the off-autocorrelation are not necessarily extracted from the "signal" buried in noise. The shapes of the on- and off-histograms are affected by artifacts introduced by the analysis procedure and depend on the signal-to-noise ratio and the overall fluorescence intensity. For experimental data where the background intensity is not constant over time we consider change point analysis to be more accurate. When using change point analysis for data of the enzyme α-chymotrypsin, no characteristics of dynamic disorder were found. In light of these results, dynamic disorder might not be a general characteristic of enzymatic reactions.
This research was jointly conducted in collaboration with the research group led by Prof. Johan Hofkens (Department of Chemistry, Katholieke Universiteit Leuven, Leuven, Belgium) and Prof. Kerstin Blank (Institute for Molecules and Materials, Radboud University, Nijmegen, The Netherlands). The outcome of our research was published in ‘ACS nano’ on 24th January 2012 (Japan time) and was a highlighted article in ACS nano, volume 6, issue 1.
Figure 1: (Left) Single molecule enzyme experiments detect the catalytic turnover of a single enzyme in terms of fluorogenic substrates. (Right) Data are collected as a time series of photon arrival time. A burst of photon indicates the completion of one catalytic cycle and the starting of the next one. The current work scrutinizes the possible artifacts in the catalytic turnover statistics due to inappropriate data analysis methodologies.