Mechanistic Investigation into Saccharomyces cerevisiae Ferulic Acid Decarboxylase, a Prenylated FMN-dependent Decarboxylase

Abstract

The UbiD family of decarboxylase enzymes uses a recently discovered, prenylated flavin mononucleotide (prFMN) as a cofactor. The best-studied UbiD enzyme is ferulic acid decarboxylase (FDC) from yeast. FDC catalyzes the reversible decarboxylation of phenylacrylic acids into their corresponding styrenes. Prior studies of FDC suggest that the mechanism involves an unprecedented 1,3-dipolar cycloaddition in which cycloadducts are formed between prFMN and substrate. Previous mechanistic studies have proposed that the rate-determining step of the reaction is cycloelimination of the styrene-prFMN cycloadduct instead of decarboxylation. However, prior to my studies, no substrate-prFMN adducts had been characterized and the pre-steady state kinetics of FDC had not been investigated. This dissertation describes the characterization of the reaction catalyzed by FDC using a combination of U.V.-visible spectroscopy, native MS, and stopped-flow absorption spectroscopy. Native MS experiments established that the styrene-prFMN cycloadduct accumulates in FDC upon reaction with either phenylacrylic acid or styrene, providing additional support that the cycloelimination step is rate-limiting. To gain further insight into the formation and breakdown of reaction intermediates, the pre-steady state kinetics of FDC reacting with both phenylacrylate and styrene were investigated. FDC exhibited surprisingly complex kinetic behavior. Reaction with either substrate resulted in loss of the prFMN iminium species monitored at 380 nm and the appearance of a broad peak at 460 nm, due to the styrene-prFMN cycloadduct. For the reaction of phenylacrylate, up to four kinetic phases could be discerned. The concentration-dependence of the amplitudes of the slower phases were directly related to substrate concentration, whereas the faster phases were relatively unaffected. This kinetic behavior suggested a half-of-sites reactivity model, a form of negative cooperativity. Global analysis of these data supported a model in which one active site of the FDC homodimer reacts rapidly with phenylacrylate in a kinetically competent manner (i.e., faster than kcat = 11.3 s-1), whereas the second active site reacts too slowly to be kinetically competent. However, a more extensive series of measurements using both phenylacrylate and styrene supports a different model. The reaction of styrene, which forms a dead-end complex, is described by three phases and is consistent with half-of-sites reactivity. For both substrates, the slower phases are no longer apparent when substrate:enzyme ratio is below 0.5, indicating one active site has a much higher affinity than the other. Although the data proved too complex to model quantitatively, they can be explained qualitatively by a “two-stroke” kinetic model in which negative cooperativity arises through catalytically relevant domain motions that have been recently proposed in the UbiD family. In this model, the high-affinity active site binds phenylacrylate and catalyzes decarboxylation to form the styrene-prFMN cycloadduct. Next, the low-affinity active site binds phenylacrylate, resulting in a conformational change that switches the affinities of the active sites. This conversion allows the energetically favorable cycloaddition and decarboxylation of phenylacrylate in the second site to be coupled to the rate-limiting cycloelimination step that releases styrene in the first site. The model is supported by the observation that the low-affinity site reacts with phenylacrylate with an observed rate constant, kobs ~8 s-1, similar to kcat. While additional experiments are necessary to fully understand the kinetics of FDC and determine the how it performs catalysis, this dissertation describes the first pre-steady state kinetic investigations into a prFMN-dependent enzyme and provides a template for future kinetic studies.