Towards next-generation enzymatic generators of hydrogen peroxide in quantitative redox biology

Thesis: Sc. D., Massachusetts Institute of Technology, Department of Chemical Engineering, June 2015.Cataloged from PDF version of thesis. "May 2015."Includes bibliographical references (pages 126-140).Hydrogen peroxide (H₂O₂) is a natural byproduct of cellular metabolism that has also been implicated in numerous biological processes, including the respiratory burst, proliferation, apoptosis, and cellular signaling. H₂O₂ has been well studied using methodologies to both measure and perturb H₂O₂ levels inside and outside cells. To perturb H₂O₂ levels, researchers have historically used bolus addition to cell culture or stimulation and inhibition of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase. However, these methodologies add conflating variables of extracellular H₂O₂, a gradient between extracellular and intracellular species, and production of superoxide (02-) as an intermediate, complicating interpretation of resulting biological effects. Furthermore, bolus addition in particular adds H₂O₂ in nonphysiological amounts, which may result in effects not seen when H₂O₂ is produced endogenously during events in which H₂O₂ has been implicated. To more accurately mimic physiological production of H₂O₂, researchers have recently turned to soluble, localizable enzymes, including glucose oxidase (GOX), xanthine oxidase (XO), and D-amino acid oxidase (DAAO). GOX, modulated by the H202 scavenger catalase, has been primarily used for extracellular generation, as has been XO; neither can be used effectively inside the cell because of GOX's requirement of a valuable metabolite, glucose, and XO's promiscuous activity on a variety of substrates and production of O₂ in addition to H₂O₂ . DAAO has been genetically encoded and used for intracellular H₂O₂ production in numerous studies; however, its requirement of exogenous substrate, typically D-alanine, and production of byproducts ammonia and [alpha]-keto acid may still introduce conflating effects. The objective of this thesis was to develop criteria for an ideal H₂O₂ generator and methodologies to engineer enzymes that meet those criteria. An ideal enzymatic H₂O₂ generator would enable meaningful perturbations to H₂O₂ levels and enable its kinetic production and steady state concentration to be quantitatively linked with signaling events and phenotypes. We first describe the criteria of an ideal H₂O₂ generator and use current kinetic parameters and concentrations of enzymes involved in H₂O₂ scavenging in HeLa cells to determine the production rate of H₂O₂ required to overcome the cell's antioxidant capacity, which constitutes one of the criteria of an ideal generator. To develop a methodology to engineer H₂O₂-generating enzymes, we sought to use an H₂O₂ sensor and Escherichia coli as a platform. Thus, in the first aim, we describe use of HyPer, a proteinaceous H₂O₂ sensor, in E. coli when H₂O₂ is added in bolus. We demonstrated that experimental parameters typically not reported, including amount of H₂O₂ per cell, cell density, E. coli strain, and timing of measurement, can significantly impact the signal. We also showed that the sensor's signal lags behind the actual amount of H₂O₂ remaining in culture during diffusion into and scavenging by E. coli, making HyPer a reversible, rather than real-time, sensor. We also generated dose-response curves and fitted these to the Hill equation, acquiring parameters that enable meaningful comparisons of the signal across studies, including dynamic range, signal-to-noise ratio, and half saturation constant. This new framework for characterizing HyPer's signal in E. coli is relevant not only to the respiratory burst, in which H₂O₂ and other related reactive species are generated to destroy invasive pathogens, but also to our work in devising a new methodology for engineering enzymes with higher H₂O₂ production. After developing a protocol to use HyPer in E. coli, we co-expressed the sensor with cytochrome P450 BM3 in E. coli to demonstrate the efficacy of a novel whole-cell screen for H₂O₂ production. We chose P450 BM3 as the target enzyme because of its satisfaction of all criteria of an ideal H₂O₂ generator except total activity in mammalian cells. We chose a co-expression scheme that minimizes variability in expression of the sensor, to avoid obfuscating interpretation of the signal. We then demonstrated that a higher signal is attained when HyPer is co-expressed with 1401P, a P450 BM3 variant known to produce H₂O₂ at a higher rate versus the wild-type enzyme. Finally, we applied a directed evolution approach to generate a library of random P450 BM3 variants and used the screen to find novel variants with enhanced H₂O₂ production, confirming the screen's efficacy. The screen significantly improves upon previous methodologies for enhancing H₂O₂ production by avoiding the need to lyse the cells and add extra reagents, reducing the amount of time required for each round of evolution. We expect the screen to find use in finding and optimizing both candidate enzymes that meet the criteria of an ideal H₂O₂ generator, as well as enzymes that produce a valuable byproduct alongside H₂O₂ Finally, we note the importance of localization of H₂O₂ generation in transducing its effects. While the current kinetic model allows calculation of the H₂O₂ production rate required to overcome the cell's antioxidant capacity, it does not account for localization and resulting spatial variations in the concentration of intracellular H₂O₂ . In the last section, we describe development of a transport model that predicts H₂O₂ concentration profiles inside the cell. We found that the kinetic model can be approximated by accounting for only peroxiredoxin as the sole H₂O₂ scavenger that does not become depleted by H₂O₂ oxidation. We thus reduced the model to a single equation, allowing us to use the finite Fourier Transform (FFT) method to develop an analytical transport model. This should enable feedback between theory and experiment, allowing us to refine the model parameters used to determine the H₂O₂ production rate necessary for an ideal generator. Together, these findings advance the field of redox biology by laying the methodologies and framework for engineering an ideal enzymatic H₂O₂ generator, the development of which should enable physiologically meaningful perturbations of H₂O₂ levels inside the cell.by Joseph Benigno Lim.Sc. D