Extracting information on reaction intermediates from single-molecule data resulting from parallel reaction mechanisms.

Abstract

Enzymes can increase the rate of a chemical reaction by stabilizing one or more reaction intermediates. Therefore, a full understanding of how a given enzyme works requires awareness of the various reaction intermediates involved in the reaction mechanism. Single-molecule studies of enzyme activity have become an increasingly common tool used to gain insight into enzyme reaction mechanisms. At the single-molecule level, product formation by any enzyme is a stochastic process. Accordingly, there is a great deal of variation in the amount of time it takes for each molecule of product to be generated, but the shape of the probability distribution for event waiting times contains information about the underlying mechanism. With adequate statistics, it is possible to determine the number of reaction intermediates for a process and their average lifetime from the experimental dwell-time distribution, even though the intermediates cannot be directly observed. However, when product can be formed via more than one pathway, the dwell-time distribution will be a mixture of the distributions for the different pathways. We use simulated data and blinded fitting to demonstrate that it is possible to extract useful information about parallel reaction pathways from mixed data distributions. By considering the fractional error of each parameter recovered, we assess what fitting methods are most successful, and determine what conditions adversely impact accuracy. We further explore the impact of heterogeneity among the lifetimes of intermediates, naturally occurring variation among enzymes, and measurement error on our ability to accurately characterize two parallel reaction mechanisms producing the same product.