Sulfur-containing compounds such as thiamin, biotin, molybdopterin, lipoic acid, and [Fe-S] clusters are essential for life. Sulfurtransferases are present in eukaryotes, eubacteria, and archaea and are believed to play important roles in mobilizing sulfur necessary for biosynthesis of these compounds and for normal cellular functions. The rhodanese homology domain is a ubiquitous structural module containing a characteristic active site cysteine residue. Some proteins containing a rhodanese domain display thiosulfate:cyanide sulfurtransferase activity in vitro. However, the physiological functions of rhodaneses remain largely unknown. YrkF, the first rhodanese to be characterized from Bacillus subtilis, is a unique protein containing two domains, an N-terminal Ccd1 domain and a C-terminal rhodanese domain. Ccd1 (conserved cysteine domain 1) is a ubiquitous structural module characterized by a Cys-Pro-X-Pro sequence motif. Thus, YrkF contains two cysteine residues (Cys¹⁵ and Cys¹⁴⁹), one in each domain. Biochemical, genetic, and bioinformatic approaches were used in order to characterize YrkF. First, YrkF was overexpressed and assayed for rhodanese activity to show that the protein is a functional rhodanese. A variant protein, YrkF^C15A, containing a cysteine to alanine substitution in the Ccd1 domain was created to determine if the Ccd1 cysteine is essential for rhodanese activity. The variant protein was overexpressed and rhodanese assays showed that YrkF^C15A is also a functional rhodanese. Inherent structural and catalytic differences were observed when comparing YrkF and YrkF^C15A, which may reflect the importance of the Ccd1 cysteine residue to normal enzymatic function and structural stability. Initial kinetic studies identified differences in activity between YrkF and YrkF^C15A. Cross-linking experiments showed a propensity for the formation of inter- and intramolecular disulfide bonds between the two cysteine residues and indicated that Cys¹⁵ and Cys¹⁴⁹ are located near one another in the 3-dimensional structure of the protein. Analysis of the proteins by mass spectrometry suggested YrkF contains a stable persulfide sulfur, whereas YrkF^C15A showed no evidence of a stable persulfide sulfur and was prone to oxidation and other active site modifications. A homology model of YrkF was created using structures of a rhodanese homolog and a Ccd1 homolog as templates. The model was used to predict the structure of YrkF based on the results of the cross-linking experiments. A strain containing a yrkF chromosomal deletion could be constructed, indicating YrkF is not essential for survival. Phenotypic analysis of the yrkF mutant revealed that YrkF is not needed for biosynthesis of sulfur-containing cofactors (thiamin, biotin, molybdopterin, or lipoic acid) or amino acids. The characterization of YrkF could lead to the discovery of novel physiological roles for rhodaneses and may give insight into possible roles for the Ccd1 module. / Master of Science

Les 3-mercaptopyruvate sulfurtransférases (3-MST) et les thiosulfate sulfurtransférases (TST) sont des enzymes ubiquitaires de la famille des thioltransférases à domaine rhodanèse qui catalysent le transfert d’un atome de soufre d’un substrat donneur vers un substrat accepteur via un intermédiaire Cys-persulfure. Les 3-MST sont impliquées dans la formation de sulfure d’hydrogène (H2S), un gazotransmetteur toxique à forte concentration, alors que les TST interviendraient dans son élimination. L’objectif de mon projet était de décrypter les mécanismes moléculaires impliquant ces thioltransférases afin de mieux comprendre leurs rôles physiologiques. Pour cela, le mécanisme catalytique et les spécificités de substrats des enzymes humaines (3-MST, TSTD1 et Rhodanèse) et d’Escherichia coli (3-MST et GlpE) ont été caractérisés grâce à la mise au point de méthodes spécifiques permettant l’étude de chacune des étapes du mécanisme (fluorescence, stopped-flow, sonde H2S) et par une étude des relations structure-fonction menée en collaboration pour les aspects chimie théorique et cristallographie RX. J’ai montré que le site actif de ces enzymes est adapté à la catalyse d’un transfert de S0 à partir de composés soufrés non activés. De plus, le mécanisme de formation de l’intermédiaire persulfure ne dépend pas de l’enzyme mais du substrat donneur. En effet, la rupture de la liaison C-S du 3-mercaptopyruvate requiert la déprotonation des fonctions thiols du substrat et de la Cys essentielle, fonction assurée par la boucle catalytique CysX5 qui constitue un véritable site de reconnaissance thiolate, et l’intervention concomitante d’une molécule d’eau comme catalyseur acide. En présence de thiosulfate, hormis l’activation de la Cys seule la neutralisation des charges négatives du substrat est indispensable à la réaction de transfert de soufre. Enfin, et de façon inattendue, la 3-MST humaine pourrait être impliquée dans l’élimination cytosolique du sulfite, un composé toxique pour les cellules. Quant aux deux TST mitochondriales humaines, elles pourraient intervenir à la fois dans la signalisation cellulaire H2S-dépendante, via la production d’espèces polysulfure, et dans l’élimination d’H2S / 3-mercaptopyruvate sulfurtransferases (3-MSTs) and thiosulfate sulfurtransferases (TSTs) are ubiquitous enzymes that belong to the rhodanese sulfurtransferase family and catalyze the transfer of a sulfur atom from a donor to an acceptor substrate via a cysteine-persulfide intermediate. While 3-MSTs are involved in the biogenesis of hydrogen sulfide (H2S), a gasotransmitter known to be toxic at high concentration, TSTs are likely responsible of its degradation. My project mainly focused on deciphering the sulfurtransferase-dependent molecular mechanisms to better define their physiological functions. To address these questions, their catalytic mechanisms and substrate specificities were investigated. This was achieved through the development of kinetic approaches (fluorescence, stopped-flow, H2S specific probe) to study each step of the reaction catalyzed by human (3-MST, TSTD1 and Rhodanese) and Escherichia coli (3-MST, GlpE) enzymes and structure-function relationship studies performed in collaboration for the theoretical chemistry and the X-ray crystallography parts. Here, I show that the active site of these enzymes is optimized to perform an efficient S0 transfer from non-activated sulfur compounds. Moreover, the mechanisms leading to formation of the persulfide intermediate do not depend on the enzyme but rather on the donor substrate. Indeed, the cleavage of the carbon-sulfur bond of 3-mercaptopyruvate critically depends on the CysX5 catalytic loop acting as a thiolate hole to favor the deprotonation of the essential Cys and the substrate, and on a water-mediated protonation step. In the presence of thiosulfate, the Cys activation mode remains unchanged and the reaction of sulfur transfer is only driven by the neutralization of the negative charges of the substrate. In addition, we propose a new physiological function for the human 3-MST in the cytoplasmic elimination of sulfite, a toxic compound for the cells. Finally, the two human mitochondrial TSTs are likely to be involved in the H2S-mediated cellular signaling, through the formation of polysulfide entities, but also in H2S catabolism

The Ni-Fe-S C-cluster of carbon monoxide dehydrogenases (CODH), which catalyzes the reversible oxidation of CO to CO2, can be stabilized in four redox states: Cox, Cred1, Cint, and Cred2. The best-supported mechanism of catalysis involves a one-electron reductive activation of Cox to Cred1 and a catalytic cycle in which Cred1 binds and oxidizes CO, forming Cred2 and releasing CO2. Recently reported experiments appear to have disqualified this mechanism, as activation was concluded to require reduction to a C-cluster state more reduced than Cred1. The results presented in this dissertation suggest that the activation potential was milder than that required to reduce these clusters. The results support a mechanism in which Cred1 is the form of the cluster that reacts with CO. The structure of the active-site C-cluster in CO dehydrogenase from Carboxydothermus hydrogenoformans (CODHCh) includes a ??2-sulfide ion bridged to the Ni and unique Fe, while the same cluster in enzymes from Rhodospirillum rubrum (CODHRr) and Moorella thermoacetica (CODHMt) lack this ion. This difference was investigated by exploring effects of sulfide on activity and spectral properties. Sulfide partially inhibited CO oxidation activities of CODHRr and CODHMt. Adding sulfide to CODHMt in the Cred1 state caused the gav = 1.82 Electron Paramagnetic Resonance spectroscopy (EPR) signal to decline and new features to appear. Sulfide did not affect the gav = 1.86 signal from the Cred2 state. A model was developed in which sulfide binds reversibly to Cred1, inhibiting catalysis. The results also suggest that the substrate hydroxyl group bridges the Ni and unique Fe. A cysteine residue recently found to form a persulfide bond with the C-cluster was characterized. The Ser mutant of the persulfide-forming Cys316 was inactive and displayed no C-cluster EPR signals. Electronic absorption and metal analysis suggest that the C-cluster is absent in this mutant. The persulfide bond appears to be essential for either the assembly or stability of the C-cluster, and/or for eliciting the redox chemistry of the C-cluster required for catalytic activity.

The Ni-Fe-S C-cluster of carbon monoxide dehydrogenases (CODH), which catalyzes the reversible oxidation of CO to CO2, can be stabilized in four redox states: Cox, Cred1, Cint, and Cred2. The best-supported mechanism of catalysis involves a one-electron reductive activation of Cox to Cred1 and a catalytic cycle in which Cred1 binds and oxidizes CO, forming Cred2 and releasing CO2. Recently reported experiments appear to have disqualified this mechanism, as activation was concluded to require reduction to a C-cluster state more reduced than Cred1. The results presented in this dissertation suggest that the activation potential was milder than that required to reduce these clusters. The results support a mechanism in which Cred1 is the form of the cluster that reacts with CO. The structure of the active-site C-cluster in CO dehydrogenase from Carboxydothermus hydrogenoformans (CODHCh) includes a ??2-sulfide ion bridged to the Ni and unique Fe, while the same cluster in enzymes from Rhodospirillum rubrum (CODHRr) and Moorella thermoacetica (CODHMt) lack this ion. This difference was investigated by exploring effects of sulfide on activity and spectral properties. Sulfide partially inhibited CO oxidation activities of CODHRr and CODHMt. Adding sulfide to CODHMt in the Cred1 state caused the gav = 1.82 Electron Paramagnetic Resonance spectroscopy (EPR) signal to decline and new features to appear. Sulfide did not affect the gav = 1.86 signal from the Cred2 state. A model was developed in which sulfide binds reversibly to Cred1, inhibiting catalysis. The results also suggest that the substrate hydroxyl group bridges the Ni and unique Fe. A cysteine residue recently found to form a persulfide bond with the C-cluster was characterized. The Ser mutant of the persulfide-forming Cys316 was inactive and displayed no C-cluster EPR signals. Electronic absorption and metal analysis suggest that the C-cluster is absent in this mutant. The persulfide bond appears to be essential for either the assembly or stability of the C-cluster, and/or for eliciting the redox chemistry of the C-cluster required for catalytic activity.

Hydrogen sulfide (H₂S) was recognized as a critical signaling molecule in mammals nearly two decades ago. Since this discovery biologists and chemists have worked in concert to demonstrate the physiological roles of H₂S as well as the therapeutic benefit of exogenous H₂S delivery. As the understanding of H₂S physiology has increased, the role(s) of other sulfur-containing molecules as potential players in cellular signaling and redox homeostasis has begun to emerge. This creates new and exciting challenges for chemists to synthesize compounds that release a signaling compound in response to specific, biologically relevant stimuli. Preparation of these signaling compound donor molecules will facilitate further elucidation of the complex chemical interplay within mammalian cells. To this end we report on two systems for the sustained release of H₂S, as well as other sulfur signaling molecules. The first system discussed is based on the N-thiocarboxyanhydride (NTA) motif. NTAs were demonstrated to release carbonyl sulfide (COS), a potential sulfur signaling molecule, in response to biologically available nucleophiles. The released COS is shown to be rapidly converted to H₂S in the presence of the ubiquitous enzyme carbonic anhydrase (CA). A synthetic route that affords NTAs with reactive functionalities was devised and the functional "parent" NTAs were successfully conjugated to a variety of substrates, ranging from small molecules to polymers. These functional NTAs provide a platform from which a library of NTA-based COS/H₂S may be readily prepared convergently in an effort to move towards H₂S-releasing drug and polymer conjugates. Additionally, preliminary in vitro cytotoxicity studies indicate that NTAs are noncytotoxic at concentrations above 100 µM. The second system discussed in this dissertation leverages the 1,6-benzyl elimination reaction (or self-immolative reaction) to facilitate the release of a persulfide (R–SSH) from a small molecule prodrug platform as well as a separate system that releases COS/H₂S from a polymer. The self-immolative persulfide prodrug was designed to be responsive to reactive oxygen species (ROS) and demonstrates efficacy as an antioxidant in vitro. Furthermore, the polymeric COS/H₂S self-immolative system was designed to respond to reducing agents, including H₂S itself, and shows promise as a H₂S signal amplification platform. / Doctor of Philosophy / Hydrogen sulfide (H₂S) has long been recognized as a malodorous and toxic byproduct of industrial chemical processes. However, the discovery of H₂S as a key signaling molecule in mammals has drastically shifted the paradigm of H₂S research over the last two decades. Research into the production and roles of H₂S in the body is ongoing, but has pointed to the implication of changes in H₂S production to the onset of a variety of disease states, including cardiovascular disease and Alzheimer’s. As alterations in the body’s production of H₂S have been correlated to certain disease states, collaborative research efforts among biologists and chemists have demonstrated the utility of H₂S-based therapeutics in helping to alleviate these disease states. Our understanding of the roles of H₂S in the body, and potential benefits derived from H₂S-releasing drugs, can only continue to advance with the development and improvement of H₂S releasing compounds. The first portion of this dissertation focuses on the synthesis of a new class of H₂S-releasing compounds, termed N-thiocarboxyanhydrides (NTAs). NTAs release H₂S through an intermediate sulfur-containing molecule, carbonyl sulfide (COS), which may have signaling properties independent of H₂S. The COS that is released from the NTAs is rapidly converted to H₂S by the action of the ubiquitous enzyme carbonic anhydrase. A variety of functional NTAs were synthesized, which in turn were used to prepare a small library of NTA-based COS/H₂S releasing compounds. This work informs the preparation of H₂S-drug or H₂S-polymer conjugates. The second portion of this dissertation examines a class of compounds broadly termed self-immolative prodrugs. The self-immolative prodrug platform was leveraged to release H₂S, or persulfides (R–SSH), another class of sulfur-containing molecules of biological interest. The self-immolative persulfide prodrug system was designed to be responsive to reactive oxygen species (ROS), a harmful cellular byproduct. The persulfide donor was successful in mitigating the harmful effects of ROS in heart cells. Independently, a polymeric self-immolative H₂S releasing system was designed to depolymerize in the presence of H₂S, resulting in the generation of 6-8-fold excess of H₂S upon depolymerization. We envision the self-immolative H₂S-releasing polymer will show promise in biological applications where a vast excess of H₂S is needed rapidly.

L’ataxie de Friedreich est une maladie neurodégénérative sévère causée par un défaut d’expression de la frataxine (FXN), une petite protéine mitochondriale impliquée dans la biogenèse des centres fer-soufre (Fe/S), des groupement prosthétiques aux fonctions cellulaires essentielles. Chez les mammifères, il a été montré que la frataxine stimule la synthèse in vitro de centres Fe/S sur la protéine d’échaffaudage ISCU, grâce à l’augmentation de la production d’ions sulfures par le complexe NFS1-ISD11-ISCU. Cependant, le mécanisme par lequel la frataxine active la biogenèse des centres Fe/S n’a pas encore été défini. Nous avons étudié les effets de FXN sur les cinétiques de formation et de réduction des persulfures, des intermédiaires clés de la production d’ions sulfures, générés par la cystéiene désulfurase NFS1, à l’aide d’un test de détection des persulfures basé sur l’utilisation de composés synthétiques peptide-maléimide et de la spectrométrie de masse. Nous avons montré que FXN active deux réactions très similaires : la réduction du persulfure de NFS1 par des réducteurs à thiols comme le DTT, la L-cystéine et le glutathion et le transfert de soufre de NFS1 vers ISCU, conduisant à l’accumulation de persulfure sur la cystéine C104 d’ISCU. Nous avons constaté que la vitesse de réduction du persulfure d’ISCU par les thiols n’est pas affectée en présence de FXN et que ce persulfure est réduit plus lentement que celui de NFS1. Nous avons corrélé l’activation par FXN de la réduction du persulfure de NFS1 par les thiols à une stimulation de l’assemblage d’un centre Fe/S sur ISCU. Dans nos conditions expérimentales, l’atome de soufre du persulfure d’ISCU n’est pas incorporé dans le centre Fe/S synthétisé, mais nos résultats ne permettent pas d’exclure que ce persulfure puisse être réduit par une réductase dédiée, encore non identifiée. L’ensemble de nos données indiquent que le rôle de la frataxine est de contrôler la réduction du persulfure de NFS1, en augmentant les vitesses de transfert de soufre vers ISCU et de réduction du persulfure de NFS1 par les thiols. / Friedreich ataxia is a severe neurodegenerative disease caused by reduced expression of frataxin (FXN), a small mitochondrial protein involved in iron-sulfur (Fe/S) cluster biogenesis which are prostetic groups with essential cellular functions. It has been shown in vitro that mammalian FXN activates Fe/S cluster synthesis on the scaffold protein ISCU, by rising up suflide ion production by NFS1-ISD11-ISCU complex. However, the mechanism by which frataxin stimulates Fe/S cluster biogenesis has not been yet defined. We have studied the effect of FXN on the kinetics of formation and reduction of persulfides that are key intermediates of sulfide ion production generated by NFS1, using mass spectrometry and a new detection assay for persulfide based on gel-mobility shift following alkylation by maleimide-peptide compounds. We demonstrate that frataxin activates two similar reactions : sulfur transfer from cysteine desulfurase NFS1 to ISCU leading to accumulation of a persulfide on ISCUcysteine C104 and reduction of NFS1 persulfide by thiol reducers such as DTT, L-cysteine and glutathion. We have observed that FXN does not stimulate the rate of ISCU persulfide reduction by thiols and that this persulfide is reduced much more slowly than NFS1 persulfide. We have then correlated the reduction of NFS1 persulfide with Fe/S cluster assembly. Under our experimental conditions, the sulfur from ISCU persulfide is not incorporated into the Fe/S cluster. However, we cannot exclude that an as yet not identfiied reductase could reduces ISCU persulfide and trigger Fe/S cluster assembly. Overall, our data point to a regulatory function of FXN as an enhancer of persulfide reduction, stimulating the rates of sulfur transfer to ISCU and NFS1 persulfide.