Cancer is a disease characterized by complexity and unpredictability. Consequently, its treatment is difficult and all too often unsuccessful. Almost all cancers are treated with some combination of the traditional anti-cancer armamentarium: surgery, chemotherapy, and radiotherapy. Recently, however, gene therapy has emerged as a promising addition to this existing repertoire. Its application as a single agent, or in combination with other anti-cancer treatments is proving successful in both pre-clinical and clinical settings. In this work I have investigated the combination of a conventional chemotherapy drug, cisplatin, with a type of cancer gene therapy known as cytosine deaminase + 5-fluorocytosine suicide gene therapy. Suicide gene therapy is the intracellular conversion of non-toxic prodrug to its active form by a metabolic enzyme of non-mammalian origin. There are many established enzyme/prodrug combinations, but here the bacterial enzyme cytosine dearninase (CDA) was used to convert inert 5-fluorocytosine (5FC) to highly toxic 5-fluorouracil (5FU). Of the various vector systems for therapeutic gene delivery, adenoviral (Ad) vectors have proven particularly suitable for application to cancer. This work used a first generation adenovirus type 5 vector expressing the enzyme cytosine deaminase (AdCDA) cloned from E. coli. The combination of AdCDA/5FC with cisplatin was chosen because the combination of 5FU and cisplatin, both of which are used extensively in cancer treatment, has proven effective clinically and demonstrates synergy in vitro. This combination was evaluated in murine mammary carcinoma MTIA2 cells, human colorectal carcinoma HT29 cells, HT29pl4 cells, the photofrin resistant sub-line of HT29 cells, and murine melanoma Bl6/FIO cells. The classical clonogenic assay was used to evaluate this combination treatment since it provides an accurate indication of the effectiveness a cancer treatment will have in vivo. AdCDA infected MTIA2, HT29, and HT29pl4 cell lines exhibited a dose response to increasing concentrations of SFC that was significantly different from control vector infected cells. Similarly, uninfected cells demonstrated a dose response to increasing concentrations of cisplatin. The effect of the combination on clonogenic survival, administered in the sequence of a 48 h exposure to SFC followed by 1 h exposure to cisplatin, was greater than additive compared to the effect of the two treatments alone. F10 cells exhibited a dose response to increasing concentrations of cisplatin. However, it could not be shown reproducibly that AdCDA infected FlO cells exhibited a dose response to SFC that differed significantly from control vector infected cells. Work with the FlO cells was inconclusive regarding the combination treatment, but it rendered information regarding the sensitivity of these cells to what is hypothesized to be an unidentified component present in some preparations of 5FC. Evaluation of this treatment in vivo, using both murine and human tumor cell lines, will further define the potential of AdCDA/5FC + cisplatin as a clinically relevant cancer treatment. / Thesis / Master of Science (MSc)

In this study, we synthesized small-molecule, water-soluble, fluorescein-containing ureido compounds 6 and 8 as target receptors for cytosine free bases and then investigated the binding of cytosine free bases with the receptors using 15N NMR spectroscopy and partially labeled cytosine-2,4-13C-1,3,4-15N-cytosine. Binding with the receptor 6a (the disodium form of 6) caused the chemical shift of the nitrogen atom of the amino group of cytosine to move downfield; binding of the receptor 8a (the disodium form of 8), which is possessing no corresponding aryl nitrogen atom, had no effect on this signal. Fluorescence spectroscopy revealed that binding of cytosine and its derivatives led to quenching of the fluorescence of receptor 6a; in contrast, the quenching of receptor 8a was only slightly affected by cytosine. Because the fluorescence of 6a was not quenched by either deoxycytidine or uracil, it appears that this receptor is a specific for cytosine among the DNA bases. We used the fluorescence of 6a to measure the apparent binding constants for various cytosine derivatives, including the anticancer prodrug 5-fluorocytosine. Receptor 6a is the first small-molecule, water-soluble fluorescent receptor for the specific binding of cytosine free bases in aqueous solution.

O pró-fármaco antimetabólito 5-Fluorocitosina (5-FC) foi investigado no campo da Engenharia de Cristais (EC) segundo a abordagem de cocristais farmacêuticos, a fim de modular sua baixa estabilidade física em ambientes úmidos, o que leva à incorporação irreversível de uma molécula de água a nível estrutural em condições de armazenamento variáveis. A forma anidra da 5-FC é um análogo fluorado da citosina muito bem conhecido por sua atividade antifúngica e com isto tornou-se um dos insumos farmacêuticos ativos (IFAs) mais utilizados para o tratamento anticâncer direcionado por meio de terapia gênica. Neste estudo, novos cocristais de 5-FC foram obtidos a partir da reação supramolecular deste IFA com o IFA tuberculostático Isoniazida (INH), bem como com outros três coformadores listados como não tóxicos: cafeína (CAF), ácido p-aminobenzóico (PABA) e ácido caprílico (CA). As amostras foram caracterizadas por difração de raios X em monocristal e policristal (DRXM e DRXP), espectroscopia na região do infravermelho (IV) e espalhamento Raman (Raman); assim como pelas técnicas de análise termogravimétrica (TG), calorimetria exploratória diferencial (CED) e microscopia termo-óptica (MTO). A estabilidade física da 5-FC e seus respectivos cocristais foi avaliada em ambiente com aproximadamente 100% de umidade relativa e a solubilidade no equilíbrio medida em meio tamponado a pH 1,2 – mimetizando valores próximos ao do suco gástrico. Os estudos estruturais mostraram que a 5-FC é capaz de formar diferentes homo e heterossíntons que levam à formação de formas multicomponentes estáveis. Dados de IV e Raman forneceram evidências espectroscópicas sobre o envolvimento dos grupos funcionais na manutenção dos principais síntons e, por tanto, do empacotamento cristalino, confirmando assim a natureza neutra necessária para a obtenção de um cocristal. Pelas análises térmicas foi possível observar que todas as amostras apresentaram uma maior preferência pela degradação do que pela mudança da fase sólida para a líquida com o fornecimento de calor, corroborando que as ligações intermoleculares de hidrogênio que mantém estas formas sólidas são fortes. Adicionalmente, constatou-se que os perfis de solubilidade dos quatro cocristais são similares ao IFA de partida, um fármaco classificado como de classe I pelo Sistema Biofarmacêutico, exibindo, assim, alta solubilidade. A instabilidade frente à hidratação bem como sua irreversibilidade foram estudadas por DRXP à temperatura ambiente (25 °C) e por DRXP em função da temperatura (até 150 °C), respectivamente. Nos cocristais, por sua vez, nenhuma transição de fase pode ser assinalada. Deste modo, todos os cocristais de 5-FC aqui reportados mantiveram uma solubilidade aceitável e não hidrataram ou sofreram transição de fase sob condições extremas de armazenamento (estudo de estabilidade acelerada em atmosfera úmida) e muito menos ao final de 12 meses de estoque (estudo de estabilidade a longo prazo), sendo mais estáveis que o IFA 5-FC forma comercializada. Além disso, o cocristal fármaco-fármaco intitulado 5FC-INH é um potencial candidato para o tratamento concomitante de infecções fúngicas, tuberculose e câncer; principalmente de pulmão. / The prodrug antimetabolite 5-Fluorocytosine (5-FC) was investigated in the field of Crystal Engineering (CE) according to the Pharmaceutical Cocrystals approach, in order to modulate its poor physical stability in humid environments, which leads to the irreversible incorporation of a water molecule at structural level under variable storage conditions. 5-FC anhydrous form is a well-known fluorinated analog of cytosine with antifungal activity and it has become one of the most used active pharmaceutical ingredients (IFAs) for anticancer treatment directed through gene therapy. In this study, novel 5-FC cocrystals were obtained from the reaction of 5-FC with the tuberculostatic IFA Isoniazid (INH) as well as with other three coformers listed as nontoxic: caffeine (CAF), p-aminobenzoic acid (PABA) and caprylic acid (CA). The samples were characterized by single-crystal and powder X-ray diffraction (SCXRD and PXRD), spectroscopic (IR and Raman) and thermogravimetric analysis (TGA), differential scanning calorimetry (DSC) and Hot-Stage microscopy (HSM) techniques. The physical stability of 5-FC and its cocrystals were evaluated in environment with high relative humidity (approximately 100 %) and the equilibrium solubility was measured in pH 1.2 buffer media – mimicking values close to that of gastric juice. The structural studies show that the prodrug 5-FC is able to form different homo and heterosynthons that lead to the formation of stable multicomponent forms. IR and Raman data provided spectroscopic evidence on the involvement of functional groups in the maintenance of major synthons and crystal packing assembly, thereby confirming the neutral nature required to obtain a cocrystal. From the thermal analyses it was possible to observe that all the samples presented a preference for degradation instead of phase transition form solid to liquid with the heat supply, corroborating the strength of intermolecular hydrogen bonds that maintain these solid forms. Additionally, the solubility profiles were found to be similar to those of the 5-FC API raw material, a Biopharmaceutical System classified as Class I drug, exhibiting high solubility profile. The instability against hydration and its irreversibility was studied by PXRD at room temperature (25 °C) and by PXRD as a function of temperature (up to 150 °C), respectively. In the cocrystals, in turn, no phase transition was found. Thus, all 5-FC cocrystals reported maintained acceptable solubility and did not hydrate or undergo phase transition under extreme storage conditions (accelerated stability study in moist atmosphere) even at the end of 12 months of storage (long-term stability study), being more stable than the commercially available IFA 5-FC. Furthermore, the drug-drug cocrystal (5FC-INH) is a potential candidate for the treatment of concomitantly fungal infections, tuberculosis and cancer, mainly lung cancer.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Infrared multiple photon dissociation (IRMPD) spectroscopy has been coupled with electronic structure calculations in order to elucidate the structures of several small biological molecules including: uracil, 5-fluorouracil, 5-fluorocytosine, ferulic acid, and a number of their related analogs. IRMPD is a powerful technique, that when combined with electronic structure calculations can provide convincing evidence for the structural characterization of ions in the gas phase. Isomers of uracil and 5-fluorouracil (5-FU) have been characterized by calculations performed at the MP2(full)/aug-cc-pVTZ level of theory; however, infrared multiple photon dissociation spectroscopy experiments proved to be unsuccessful for these species. Geometry optimization and frequency calculations have isolated the dominant isomer(s) for neutral and deprotonated uracil and 5-fluorouracil, along with several cluster interactions involving water, methanol, ammonia, and methylamine. For both uracil and 5-FU, a single relevant neutral isomer was determined, with each isomer existing in the diketo, as opposed to the enol form. Following the deprotonation of this neutral isomer, both uracil and 5-FU were permitted to form anionic cluster ions with water, methanol, ammonia, or methylamine, and based on the relative Gibbs free energies (298 K) of the calculated isomers, relevant cluster interactions were determined. For each cluster, several sites of intramolecular interaction were found to exist; however, interaction at the site of deprotonation was the most favourable in every instance. Ionic hydrogen bond interactions have been found in several clusters formed by 5-fluorocytosine (5-FC). The chloride and trimethylammonium cluster ions, in addition to the cationic and anionic dimers have been characterized by infrared multiple photon dissociation (IRMPD) spectroscopy and electronic structure calculations performed at the B2PLYP/aug-cc-pVTZ//B3LYP/6-311+G(d,p) level of theory. IRMPD spectra in combination with calculated spectra and relative energetics have indicated, quite conclusively, that a single isomer for each 5-FC cluster that is likely being observed experimentally except in the case of the anionic dimer, in which a combination of isomers is probable. For the 5-FC-trimethylammonium cluster specifically, the calculated spectrum of the lowest energy isomer matches the experimental spectrum remarkably well. Interestingly, the cationic dimer of 5-FC was found to have a single energetically relevant isomer (Cationic-IV) in which a unique tridentate ionic hydrogen bond interaction is formed. The three sites of intramolecular ionic hydrogen bonds in this isomer interact very efficiently, leading to a significantly large calculated enthalpy of binding of 180 kJ/mol. The magnitude of the calculated binding energy for this species, in combination with the strong correlation between the simulated and IRMPD spectra, indicates that the tridentate-bound dimer is observed predominantly in experiment. Comparison of the calculated relative Gibbs free energies (298 K) for this species with several of the other isomers considered also supports the likelihood of the dominant protonated dimer existing as Cationic-IV. Protonated ferulic acid has been characterized using infrared multiple photon dissociation spectroscopy and electronic structure calculations at the B3LYP/6-311+G(d,p) level of theory. Neutral ferulic acid has been determined to undergo protonation on the carbonyl oxygen of the acid group, forming an ion of m/z 195. Due to its extensively conjugated structure, protonated ferulic acid (m/z 195) is observed to yield three stable fragment ions in IRMPD experiments. It is proposed that two parallel fragmentation pathways of protonated ferulic acid are being observed. First, proton transfer occurs from the carbonyl oxygen to the hydroxyl oxygen within the acid group, resulting in the loss of water and subsequently carbon monoxide, forming ions of m/z 177 and 149, respectively. The second proposed fragmentation pathway undergoes proton transfer from the phenolic group to the methoxy group resulting in loss of methanol and rearrangement to a five-membered ring of m/z 163. IRMPD spectra have been obtained for the ions m/z 195 and m/z 177, and anharmonic calculations have been performed on these species at the B3LYP/6-311+G(d,p) level of theory. The calculated anharmonic spectra for these ions match the experimental spectrum exceptionally well and strongly support the proposed fragmentation mechanisms.

Infrared multiple photon dissociation (IRMPD) spectroscopy has been coupled with electronic structure calculations in order to elucidate the structures of several small biological molecules including: uracil, 5-fluorouracil, 5-fluorocytosine, ferulic acid, and a number of their related analogs. IRMPD is a powerful technique, that when combined with electronic structure calculations can provide convincing evidence for the structural characterization of ions in the gas phase. Isomers of uracil and 5-fluorouracil (5-FU) have been characterized by calculations performed at the MP2(full)/aug-cc-pVTZ level of theory; however, infrared multiple photon dissociation spectroscopy experiments proved to be unsuccessful for these species. Geometry optimization and frequency calculations have isolated the dominant isomer(s) for neutral and deprotonated uracil and 5-fluorouracil, along with several cluster interactions involving water, methanol, ammonia, and methylamine. For both uracil and 5-FU, a single relevant neutral isomer was determined, with each isomer existing in the diketo, as opposed to the enol form. Following the deprotonation of this neutral isomer, both uracil and 5-FU were permitted to form anionic cluster ions with water, methanol, ammonia, or methylamine, and based on the relative Gibbs free energies (298 K) of the calculated isomers, relevant cluster interactions were determined. For each cluster, several sites of intramolecular interaction were found to exist; however, interaction at the site of deprotonation was the most favourable in every instance. Ionic hydrogen bond interactions have been found in several clusters formed by 5-fluorocytosine (5-FC). The chloride and trimethylammonium cluster ions, in addition to the cationic and anionic dimers have been characterized by infrared multiple photon dissociation (IRMPD) spectroscopy and electronic structure calculations performed at the B2PLYP/aug-cc-pVTZ//B3LYP/6-311+G(d,p) level of theory. IRMPD spectra in combination with calculated spectra and relative energetics have indicated, quite conclusively, that a single isomer for each 5-FC cluster that is likely being observed experimentally except in the case of the anionic dimer, in which a combination of isomers is probable. For the 5-FC-trimethylammonium cluster specifically, the calculated spectrum of the lowest energy isomer matches the experimental spectrum remarkably well. Interestingly, the cationic dimer of 5-FC was found to have a single energetically relevant isomer (Cationic-IV) in which a unique tridentate ionic hydrogen bond interaction is formed. The three sites of intramolecular ionic hydrogen bonds in this isomer interact very efficiently, leading to a significantly large calculated enthalpy of binding of 180 kJ/mol. The magnitude of the calculated binding energy for this species, in combination with the strong correlation between the simulated and IRMPD spectra, indicates that the tridentate-bound dimer is observed predominantly in experiment. Comparison of the calculated relative Gibbs free energies (298 K) for this species with several of the other isomers considered also supports the likelihood of the dominant protonated dimer existing as Cationic-IV. Protonated ferulic acid has been characterized using infrared multiple photon dissociation spectroscopy and electronic structure calculations at the B3LYP/6-311+G(d,p) level of theory. Neutral ferulic acid has been determined to undergo protonation on the carbonyl oxygen of the acid group, forming an ion of m/z 195. Due to its extensively conjugated structure, protonated ferulic acid (m/z 195) is observed to yield three stable fragment ions in IRMPD experiments. It is proposed that two parallel fragmentation pathways of protonated ferulic acid are being observed. First, proton transfer occurs from the carbonyl oxygen to the hydroxyl oxygen within the acid group, resulting in the loss of water and subsequently carbon monoxide, forming ions of m/z 177 and 149, respectively. The second proposed fragmentation pathway undergoes proton transfer from the phenolic group to the methoxy group resulting in loss of methanol and rearrangement to a five-membered ring of m/z 163. IRMPD spectra have been obtained for the ions m/z 195 and m/z 177, and anharmonic calculations have been performed on these species at the B3LYP/6-311+G(d,p) level of theory. The calculated anharmonic spectra for these ions match the experimental spectrum exceptionally well and strongly support the proposed fragmentation mechanisms.

Les protéines sont les produits finaux de la machinerie génétique. Elles jouent des rôles essentiels dans la définition de la structure, de l'intégrité et de la dynamique de la cellule afin de promouvoir les diverses transformations chimiques requises dans le métabolisme et dans la transmission des signaux biochimique. Nous savons que la doctrine centrale de la biologie moléculaire: un gène = un ARN messager = une protéine, est une simplification grossière du système biologique. En effet, plusieurs ARN messagers peuvent provenir d’un seul gène grâce à l’épissage alternatif. De plus, une protéine peut adopter plusieurs fonctions au courant de sa vie selon son état de modification post-traductionelle, sa conformation et son interaction avec d’autres protéines. La formation de complexes protéiques peut, en elle-même, être déterminée par l’état de modifications des protéines influencées par le contexte génétique, les compartiments subcellulaires, les conditions environmentales ou être intrinsèque à la croissance et la division cellulaire. Les complexes protéiques impliqués dans la régulation du cycle cellulaire sont particulièrement difficiles à disséquer car ils ne se forment qu’au cours de phases spécifiques du cycle cellulaire, ils sont fortement régulés par les modifications post-traductionnelles et peuvent se produire dans tous les compartiments subcellulaires. À ce jour, aucune méthode générale n’a été développée pour permettre une dissection fine de ces complexes macromoléculaires. L'objectif de cette thèse est d'établir et de démontrer une nouvelle stratégie pour disséquer les complexes protéines formés lors du cycle cellulaire de la levure Saccharomyces cerevisiae (S. cerevisiae). Dans cette thèse, je décris le développement et l'optimisation d'une stratégie simple de sélection basée sur un essai de complémentation de fragments protéiques en utilisant la cytosine déaminase de la levure comme sonde (PCA OyCD). En outre, je décris une série d'études de validation du PCA OyCD afin de l’utiliser pour disséquer les mécanismes d'activation des facteurs de transcription et des interactions protéine-protéines (IPPs) entre les régulateurs du cycle cellulaire. Une caractéristique clé du PCA OyCD est qu'il peut être utilisé pour détecter à la fois la formation et la dissociation des IPPs et émettre un signal détectable (la croissance des cellules) pour les deux types de sélections. J'ai appliqué le PCA OyCD pour disséquer les interactions entre SBF et MBF, deux facteurs de transcription clés régulant la transition de la phase G1 à la phase S. SBF et MBF sont deux facteurs de transcription hétérodimériques composés de deux sous-unités : une protéine qui peut lier directement l’ADN (Swi4 ou Mbp1, respectivement) et une protéine commune contenant un domain d’activation de la transcription appelée Swi6. J'ai appliqué le PCA OyCD afin de générer un mutant de Swi6 qui restreint ses activités transcriptionnelles à SBF, abolissant l’activité MBF. Nous avons isolé des souches portant des mutations dans le domaine C-terminal de Swi6, préalablement identifié comme responsable dans la formation de l’interaction avec Swi4 et Mbp1, et également important pour les activités de SBF et MBF. Nos résultats appuient un modèle où Swi6 subit un changement conformationnel lors de la liaison à Swi4 ou Mbp1. De plus, ce mutant de Swi6 a été utilisé pour disséquer le mécanisme de régulation de l’entrée de la cellule dans un nouveau cycle de division cellulaire appelé « START ». Nous avons constaté que le répresseur de SBF et MBF nommé Whi5 se lie directement au domaine C-terminal de Swi6. Finalement, j'ai appliqué le PCA OyCD afin de disséquer les complexes protéiques de la kinase cycline-dépendante de la levure nommé Cdk1. Cdk1 est la kinase essentielle qui régule la progression du cycle cellulaire et peut phosphoryler un grand nombre de substrats différents en s'associant à l'une des neuf protéines cycline régulatrice (Cln1-3, Clb1-6). Je décris une stratégie à haut débit, voir à une échelle génomique, visant à identifier les partenaires d'interaction de Cdk1 et d’y associer la cycline appropriée(s) requise(s) à l’observation d’une interaction en utilisant le PCA OyCD et des souches délétées pour chacune des cyclines. Mes résultats nous permettent d’identifier la phase(s) du cycle cellulaire où Cdk1 peut phosphoryler un substrat particulier et la fonction potentielle ou connue de Cdk1 pendant cette phase. Par exemple, nous avons identifié que l’interaction entre Cdk1 et la γ-tubuline (Tub4) est dépendante de Clb3. Ce résultat est conforme au rôle de Tub4 dans la nucléation et la croissance des faisceaux mitotiques émanant des centromères. Cette stratégie peut également être appliquée à l’étude d'autres IPPs qui sont contrôlées par des sous-unités régulatrices. / Proteins are the end-products of gene interpretative machinery. Proteins serve essential roles in defining the structure, integrity and dynamics of the cell and mediate most chemical transformations needed for everything from metabolic catalysis to signal transduction. We know that the central dogma of molecular biology, one gene = one mRNA = one protein is a gross simplification and that a protein may do different things depending on the form in which its mRNA was spliced, how and where it is post-translationally modified, what conformational state it may be in or finally, which other proteins it may interact with. Formation of protein complexes may, themselves, be governed by the states in which proteins are expressed in specific cells, cellular compartments or under specific conditions or dynamic phases such has growth or division. Protein complexes involved in mitotic cell cycle regulation are particularly challenging to dissect since they could only form during specific phases of the cell cycle, are highly regulated by post-translational modifications and can be found in any subcellular compartments. To date, no general methods have been developed to allow fine dissection of these protein complexes. The goal of this thesis was to establish and demonstrate a novel strategy for dissecting protein complexes regulating the budding yeast Saccharomyces cerevisiae (S. cerevisiae) mitotic cell cycle. In this thesis, I describe my development and optimization of a simple survival-selection Protein-fragment Complementation Assay using the prodrug-converting enzyme, yeast cytosine deaminase as reporter (OyCD PCA). I further describe, in a series of proof of principle studies, applications of the OyCD PCA to dissect the mechanism of transcriptional activation by key mitotic transcription factors and to dissect protein-protein interactions (PPIs) among regulators of the mitotic cell cycle. A key feature of the OyCD PCA is that it can be used to detect both formation and disruption of PPIs by virtue of having positive readouts for both assays. I applied the OyCD PCA in a strategy to dissect interactions between the key transcription factors of the G1/S phase: SBF and MBF. These two heterodimeric transcription factors are composed of, respectively, two distinct DNA-binding subunits named Swi4 and Mbp1 and a common transcription activation subunit called Swi6. I took advantage of the dual selection by OyCD PCA to engineer a specific mutant of Swi6 in order to demonstrate the rewiring of a transcriptional network. We isolated Swi6 with mutations found in its C-terminal domain previously identified for binding Swi4 and Mbp1 and important for SBF and MBF activities. Our results support a model where Swi6 undergoes a conformational change upon binding to Swi4 or Mbp1. In addition, this Swi6 mutant was used to dissect the regulatory mechanism that governs the entry of S. cerevisiae to a new round of cell division also known as START. We found that the SBF and MBF repressor Whi5 directly binds to the C-terminal domain of Swi6. Finally, I applied the OyCD PCA to dissect the yeast cyclin dependent kinase Cdk1-protein complexes. Cdk1 is the essential kinase that regulates cell cycle progression and can phosphorylate a large number of different substrates by teaming up with one of nine cyclin regulatory proteins (Cln1-3, Clb1-6). I describe a strategy to identify interaction partners of Cdk1 that can easily be scaled up for a genome-wide screen and associate the complexes with the appropriate cyclin(s) required for mediating the interaction using the OyCD PCA and deletion of the cyclin genes. My results allow us to postulate which phase(s) of the mitotic cell cycle Cdk1 may phosphorylate proteins and what function potential or known substrates of Cdk1 may take on during that phase(s). For example, we identified the interaction between Cdk1 and the γ-tubulin (Tub4) to be dependent upon Clb3, consistent with its role in mediating nucleation and growth of mitotic microtubule bundles on the spindle pole body. This strategy can also be applied to study other PPIs that are contingent upon accessory subunits.

Les protéines sont les produits finaux de la machinerie génétique. Elles jouent des rôles essentiels dans la définition de la structure, de l'intégrité et de la dynamique de la cellule afin de promouvoir les diverses transformations chimiques requises dans le métabolisme et dans la transmission des signaux biochimique. Nous savons que la doctrine centrale de la biologie moléculaire: un gène = un ARN messager = une protéine, est une simplification grossière du système biologique. En effet, plusieurs ARN messagers peuvent provenir d’un seul gène grâce à l’épissage alternatif. De plus, une protéine peut adopter plusieurs fonctions au courant de sa vie selon son état de modification post-traductionelle, sa conformation et son interaction avec d’autres protéines. La formation de complexes protéiques peut, en elle-même, être déterminée par l’état de modifications des protéines influencées par le contexte génétique, les compartiments subcellulaires, les conditions environmentales ou être intrinsèque à la croissance et la division cellulaire. Les complexes protéiques impliqués dans la régulation du cycle cellulaire sont particulièrement difficiles à disséquer car ils ne se forment qu’au cours de phases spécifiques du cycle cellulaire, ils sont fortement régulés par les modifications post-traductionnelles et peuvent se produire dans tous les compartiments subcellulaires. À ce jour, aucune méthode générale n’a été développée pour permettre une dissection fine de ces complexes macromoléculaires. L'objectif de cette thèse est d'établir et de démontrer une nouvelle stratégie pour disséquer les complexes protéines formés lors du cycle cellulaire de la levure Saccharomyces cerevisiae (S. cerevisiae). Dans cette thèse, je décris le développement et l'optimisation d'une stratégie simple de sélection basée sur un essai de complémentation de fragments protéiques en utilisant la cytosine déaminase de la levure comme sonde (PCA OyCD). En outre, je décris une série d'études de validation du PCA OyCD afin de l’utiliser pour disséquer les mécanismes d'activation des facteurs de transcription et des interactions protéine-protéines (IPPs) entre les régulateurs du cycle cellulaire. Une caractéristique clé du PCA OyCD est qu'il peut être utilisé pour détecter à la fois la formation et la dissociation des IPPs et émettre un signal détectable (la croissance des cellules) pour les deux types de sélections. J'ai appliqué le PCA OyCD pour disséquer les interactions entre SBF et MBF, deux facteurs de transcription clés régulant la transition de la phase G1 à la phase S. SBF et MBF sont deux facteurs de transcription hétérodimériques composés de deux sous-unités : une protéine qui peut lier directement l’ADN (Swi4 ou Mbp1, respectivement) et une protéine commune contenant un domain d’activation de la transcription appelée Swi6. J'ai appliqué le PCA OyCD afin de générer un mutant de Swi6 qui restreint ses activités transcriptionnelles à SBF, abolissant l’activité MBF. Nous avons isolé des souches portant des mutations dans le domaine C-terminal de Swi6, préalablement identifié comme responsable dans la formation de l’interaction avec Swi4 et Mbp1, et également important pour les activités de SBF et MBF. Nos résultats appuient un modèle où Swi6 subit un changement conformationnel lors de la liaison à Swi4 ou Mbp1. De plus, ce mutant de Swi6 a été utilisé pour disséquer le mécanisme de régulation de l’entrée de la cellule dans un nouveau cycle de division cellulaire appelé « START ». Nous avons constaté que le répresseur de SBF et MBF nommé Whi5 se lie directement au domaine C-terminal de Swi6. Finalement, j'ai appliqué le PCA OyCD afin de disséquer les complexes protéiques de la kinase cycline-dépendante de la levure nommé Cdk1. Cdk1 est la kinase essentielle qui régule la progression du cycle cellulaire et peut phosphoryler un grand nombre de substrats différents en s'associant à l'une des neuf protéines cycline régulatrice (Cln1-3, Clb1-6). Je décris une stratégie à haut débit, voir à une échelle génomique, visant à identifier les partenaires d'interaction de Cdk1 et d’y associer la cycline appropriée(s) requise(s) à l’observation d’une interaction en utilisant le PCA OyCD et des souches délétées pour chacune des cyclines. Mes résultats nous permettent d’identifier la phase(s) du cycle cellulaire où Cdk1 peut phosphoryler un substrat particulier et la fonction potentielle ou connue de Cdk1 pendant cette phase. Par exemple, nous avons identifié que l’interaction entre Cdk1 et la γ-tubuline (Tub4) est dépendante de Clb3. Ce résultat est conforme au rôle de Tub4 dans la nucléation et la croissance des faisceaux mitotiques émanant des centromères. Cette stratégie peut également être appliquée à l’étude d'autres IPPs qui sont contrôlées par des sous-unités régulatrices. / Proteins are the end-products of gene interpretative machinery. Proteins serve essential roles in defining the structure, integrity and dynamics of the cell and mediate most chemical transformations needed for everything from metabolic catalysis to signal transduction. We know that the central dogma of molecular biology, one gene = one mRNA = one protein is a gross simplification and that a protein may do different things depending on the form in which its mRNA was spliced, how and where it is post-translationally modified, what conformational state it may be in or finally, which other proteins it may interact with. Formation of protein complexes may, themselves, be governed by the states in which proteins are expressed in specific cells, cellular compartments or under specific conditions or dynamic phases such has growth or division. Protein complexes involved in mitotic cell cycle regulation are particularly challenging to dissect since they could only form during specific phases of the cell cycle, are highly regulated by post-translational modifications and can be found in any subcellular compartments. To date, no general methods have been developed to allow fine dissection of these protein complexes. The goal of this thesis was to establish and demonstrate a novel strategy for dissecting protein complexes regulating the budding yeast Saccharomyces cerevisiae (S. cerevisiae) mitotic cell cycle. In this thesis, I describe my development and optimization of a simple survival-selection Protein-fragment Complementation Assay using the prodrug-converting enzyme, yeast cytosine deaminase as reporter (OyCD PCA). I further describe, in a series of proof of principle studies, applications of the OyCD PCA to dissect the mechanism of transcriptional activation by key mitotic transcription factors and to dissect protein-protein interactions (PPIs) among regulators of the mitotic cell cycle. A key feature of the OyCD PCA is that it can be used to detect both formation and disruption of PPIs by virtue of having positive readouts for both assays. I applied the OyCD PCA in a strategy to dissect interactions between the key transcription factors of the G1/S phase: SBF and MBF. These two heterodimeric transcription factors are composed of, respectively, two distinct DNA-binding subunits named Swi4 and Mbp1 and a common transcription activation subunit called Swi6. I took advantage of the dual selection by OyCD PCA to engineer a specific mutant of Swi6 in order to demonstrate the rewiring of a transcriptional network. We isolated Swi6 with mutations found in its C-terminal domain previously identified for binding Swi4 and Mbp1 and important for SBF and MBF activities. Our results support a model where Swi6 undergoes a conformational change upon binding to Swi4 or Mbp1. In addition, this Swi6 mutant was used to dissect the regulatory mechanism that governs the entry of S. cerevisiae to a new round of cell division also known as START. We found that the SBF and MBF repressor Whi5 directly binds to the C-terminal domain of Swi6. Finally, I applied the OyCD PCA to dissect the yeast cyclin dependent kinase Cdk1-protein complexes. Cdk1 is the essential kinase that regulates cell cycle progression and can phosphorylate a large number of different substrates by teaming up with one of nine cyclin regulatory proteins (Cln1-3, Clb1-6). I describe a strategy to identify interaction partners of Cdk1 that can easily be scaled up for a genome-wide screen and associate the complexes with the appropriate cyclin(s) required for mediating the interaction using the OyCD PCA and deletion of the cyclin genes. My results allow us to postulate which phase(s) of the mitotic cell cycle Cdk1 may phosphorylate proteins and what function potential or known substrates of Cdk1 may take on during that phase(s). For example, we identified the interaction between Cdk1 and the γ-tubulin (Tub4) to be dependent upon Clb3, consistent with its role in mediating nucleation and growth of mitotic microtubule bundles on the spindle pole body. This strategy can also be applied to study other PPIs that are contingent upon accessory subunits.