Trois protéines de la famille TRIM (Motif TRIpartite), TIF1α, β (Transcriptional Intermediary Factor 1) et PML (ProMyelocytic Leukaemia¬), font l’objet de cette étude. TIF1α est connu comme un coactivateur des récepteurs nucléaires et TIF1β comme le corépresseur universel des protéines KRAB-multidoigt de zinc dont le prototype étudié ici est ZNF74. PML possède divers rôles dont le plus caractérisé est celui d’être l’organisateur principal et essentiel des PML-NBs (PML-Nuclear Bodies), des macrostructures nucléaires très dynamiques regroupant et coordonnant plus de 40 protéines. Il est à noter que la fonction de TIF1α, β et PML est régulée par une modification post-traductionnelle, la sumoylation, qui implique le couplage covalent de la petite protéine SUMO (Small Ubiquitin like MOdifier) à des lysines de ces trois protéines cibles. Cette thèse propose de développer des méthodes utilisant le BRET (Bioluminescence Resonance Energy Transfert) afin de détecter dans des cellules vivantes et en temps réel des interactions non-covalentes de protéines nucléaires mais aussi leur couplage covalent à SUMO. En effet, le BRET n’a jamais été exploré jusqu’alors pour étudier les interactions non-covalentes et covalentes de protéines nucléaires. L’étude de l’interaction de protéines transcriptionnellement actives est parfois difficile par des méthodes classiques du fait de leur grande propension à agréger (famille TRIM) ou de leur association à la matrice nucléaire (ZNF74). L’homo et l’hétérodimérisation de TIF1α, β ainsi que leur interaction avec ZNF74 sont ici testées sur des protéines entières dans des cellules vivantes de mammifères répondant aux résultats conflictuels de la littérature et démontrant que le BRET peut être avantageusement utilisé comme alternative aux essais plus classiques basés sur la transcription. Du fait de l’hétérodimérisation confirmée de TIF1α et β, le premier article présenté ouvre la possibilité d’une relation étroite entre les récepteurs nucléaires et les protéines KRAB- multidoigt de zinc. Des études précédentes ont démontré que la sumoylation de PML est impliquée dans sa dégradation induite par l’As2O3 et dépendante de RNF4, une E3 ubiquitine ligase ayant pour substrat des chaînes de SUMO (polySUMO). Dans le second article, grâce au développement d’une nouvelle application du BRET pour la détection d’interactions covalentes et non-covalentes avec SUMO (BRETSUMO), nous établissons un nouveau lien entre la sumoylation de PML et sa dégradation. Nous confirmons que le recrutement de RNF4 dépend de SUMO mais démontrons également l’implication du SBD (Sumo Binding Domain) de PML dans sa dégradation induite par l’As2O3 et/ou RNF4. De plus, nous démontrons que des sérines, au sein du SBD de PML, qui sont connues comme des cibles de phosphorylation par la voie de la kinase CK2, régulent les interactions non-covalentes de ce SBD mettant en évidence, pour la première fois, que les interactions avec un SBD peuvent dépendre d’un évènement de phosphorylation (“SBD phospho-switch”). Nos résultats nous amènent à proposer l’hypothèse que le recrutement de PML sumoylé au niveau des PML-NBs via son SBD, favorise le recrutement d’une autre activité E3 ubiquitine ligase, outre celle de RNF4, PML étant lui-même un potentiel candidat. Ceci suggère l’existence d’une nouvelle relation dynamique entre phosphorylation, sumoylation et ubiquitination de PML. Finalement, il est suggéré que PML est dégradé par deux voies différentes dépendantes de l’ubiquitine et du protéasome; la voie de CK2 et la voie de RNF4. Enfin une étude sur la sumoylation de TIF1β est également présentée en annexe. Cette étude caractérise les 6 lysines cibles de SUMO sur TIF1β et démontre que la sumoylation est nécessaire à l’activité répressive de TIF1β mais n’est pas impliquée dans son homodimérisation ou son interaction avec la boîte KRAB. La sumoylation est cependant nécessaire au recrutement d’histones déacétylases, dépendante de son homodimérisation et de l’intégrité du domaine PHD. Alors que l’on ne connaît pas de régulateur physiologique de la sumoylation outre les enzymes directement impliquées dans la machinerie de sumoylation, nous mettons en évidence que la sumoylation de TIF1β est positivement régulée par son interaction avec le domaine KRAB et suggérons que ces facteurs transcriptionnels recrutent TIF1β à l’ADN au niveau de promoteur et augmentent son activité répressive en favorisant sa sumoylation. / Three TRIM proteins (TRIpartite Motif), TIF1α, β (Transcriptional Intermediary Factor 1) and PML (ProMyelocytic Leukaemia¬), were studied in this thesis. TIF1α is a nuclear receptor coactivator and TIF1β is the universal corepressor of the KRAB-zinc finger repressor family of which, ZNF74 is studied here as a prototypic member. PML functions as a tumor suppressor and is the essential organiser of PML-NBs (PML-Nuclear Bodies) which are very dynamic nuclear macrostructures containing more than 40 proteins. The function of these three TRIM proteins is regulated by sumoylation, a post-translational modification involving the covalent linkage of SUMO (Small Ubiquitin like MOdifier) to specific targets lysine. In this thesis, we propose to develop new methods based on BRET (Bioluminescence Resonance Energy Transfer) to detect non-covalent nuclear protein interactions but also covalent linkage to SUMO in real time in living cells. To date, BRET was never used to assess non-covalent or covalent nuclear protein interactions. Studying transcriptionally active protein interactions represents a challenge by classical methods in particular when proteins have a tendency to aggregate (TRIM family) or when characterizing nuclear matrix proteins (ZNF74). In the first article, homo- and heterodimerisation of TIF1 α and β as well as their interaction with ZNF74 was assessed by BRET using full length proteins in living mammalian cells. We ascertained the heterodimerisation of TIF1α and β. Whereas ZNF74 interacts strongly with TIF1β, no interaction was detected with TIF1α. However, we unravelled the existence of ternary complexes involving ZNF74, TIF1α and TIF1β. This suggested that a mechanisms for cross-talk between nuclear receptors and KRAB-zinc finger proteins. Thus, we showed that BRET can be advantageously used as a non-transcription-based interaction system for studying transcriptionally active proteins, including nuclear matrix proteins, in living cells. Previous studies have shown that the sumoylation of PML (a tumour suppressor) is involved in its proteasome degradation that is As2O3-inducible and dependent on the polySUMO E3 ubiquitin ligase, RNF4. In the second article, we describe the development of a new application of the BRET method for the detection of covalent and non-covalent interactions with SUMO. Owing to this SUMO BRET assay, we established that the As2O3 / RNF4-mediated degradation of PML, not only depends on PML sumoylation as previously demonstrated, but also on the integrity of its SUMO binding domain. We also demonstrated that As2O3 which increases PML sumoylation, also enhances PML / RNF4 interaction. Our study revealed that most PML SBD non covalent interactions with sumoylated proteins required the phosphorylation of serines within PML SBD that were previously described as target sites for CK2 kinase and involved in PML degradation. Despites the involvement of PML SBD in RNF4-mediated degradation, these serines which function as an SBD phospho-switch, were not required for RNF4-mediated degradation. This suggested that CK2- and RNF4-mediated PML degradation represents two distinct pathways triggering PML ubiquitin / proteasome-dependent degradation. At last, our study led to the hypothesis that the recruitment of sumoylated PML at PML-Nuclear Bodies subnuclear structures via the PML SBD and / or possibly an E3 ubiquitin ligase activity other than RNF4 (PML itself being candidate) may favour PML degradation. Our study also stresses the dynamic involvement of three PML post-translational modifications, phosphorylation, sumoylation and ubiquitination in its degradation. A third article addressing the role of TIF1β sumoylation is presented in the Appendix. We characterized the 6 SUMO targets lysine of TIF1β and demonstrated that sumoylation is required for TIF1β transcriptional repressive activity. This is in part explained by the fact that TIF1β sumoylation is a pre-requisite for histone deacetylases recruitment since TIF1β repressive activity is partly dependent on histone deacetylases. We found that TIF1β sumoylation does not influence its homodimerisation or interaction with the KRAB box of KRAB zinc finger proteins recruiting TIF1β to promoters. TIF1β sumoylation is however relying on the integrity of TIF1β PHD finger and on its self-oligomerisation. Interestingly, we demonstrated that TIF1β sumoylation is positively regulated by its interaction with KRAB domain. It is thus suggested that KRAB-zinc finger proteins recruit TIF1β at DNA promoters where they trigger increase of TIF1β sumoylation and thus enhance its repressive activity.

In many scenarios, events such as singularities and transients that carry important information about a signal undergo spreading during acquisition or transmission and it is important to localize the events. For example, edges in an image, point sources in a microscopy or astronomical image are blurred by the point-spread function (PSF) of the acquisition system, while in a speech signal, the epochs corresponding to glottal closure instants are shaped by the vocal tract response. Such events can be extracted with the help of techniques that promote sparsity, which enables separation of the smooth components from the transient ones. In this thesis, we consider development of such sparsity promoting techniques. The contributions of the thesis are three-fold: (i) an auditory-motivated continuous wavelet design and representation, which helps identify singularities; (ii) a sparsity-driven deconvolution technique; and (iii) a sparsity-driven deconvolution technique for reconstruction of nite-rate-of-innovation (FRI) signals. We use the speech signal for illustrating the performance of the techniques in the first two parts and super-resolution microscopy (2-D) for the third part. In the rst part, we develop a continuous wavelet transform (CWT) starting from an auditory motivation. Wavelet analysis provides good time and frequency localization, which has made it a popular tool for time-frequency analysis of signals. The CWT is a multiresolution analysis tool that involves decomposition of a signal using a constant-Q wavelet filterbank, akin to the time-frequency analysis performed by basilar membrane in the peripheral human auditory system. This connection motivated us to develop wavelets that possess auditory localization capabilities. Gammatone functions are extensively used in the modeling of the basilar membrane, but the non-zero average of the functions poses a hurdle. We construct bona de wavelets from the Gammatone function called Gammatone wavelets and analyze their properties such as admissibility, time-bandwidth product, vanishing moments, etc.. Of particular interest is the vanishing moments property, which enables the wavelet to suppress smooth regions in a signal leading to sparsi cation. We show how this property of the Gammatone wavelets coupled with multiresolution analysis could be employed for singularity and transient detection. Using these wavelets, we also construct equivalent lterbank models and obtain cepstral feature vectors out of such a representation. We show that the Gammatone wavelet cepstral coefficients (GWCC) are effective for robust speech recognition compared with mel-frequency cepstral coefficients (MFCC). In the second part, we consider the problem of sparse blind deconvolution (SBD) starting from a signal obtained as the convolution of an unknown PSF and a sparse excitation. The BD problem is ill-posed and the goal is to employ sparsity to come up with an accurate solution. We formulate the SBD problem within a Bayesian framework. The estimation of lter and excitation involves optimization of a cost function that consists of an `2 data- fidelity term and an `p-norm (p 2 [0; 1]) regularizer, as the sparsity promoting prior. Since the `p-norm is not differentiable at the origin, we consider a smoothed version of the `p-norm as a proxy in the optimization. Apart from the regularizer being non-convex, the data term is also non-convex in the filter and excitation as they are both unknown. We optimize the non-convex cost using an alternating minimization strategy, and develop an alternating `p `2 projections algorithm (ALPA). We demonstrate convergence of the iterative algorithm and analyze in detail the role of the pseudo-inverse solution as an initialization for the ALPA and provide probabilistic bounds on its accuracy considering the presence of noise and the condition number of the linear system of equations. We also consider the case of bounded noise and derive tight tail bounds using the Hoe ding inequality. As an application, we consider the problem of blind deconvolution of speech signals. In the linear model for speech production, voiced speech is assumed to be the result of a quasi-periodic impulse train exciting a vocal-tract lter. The locations of the impulses or epochs indicate the glottal closure instants and the spacing between them the pitch. Hence, the excitation in the case of voiced speech is sparse and its deconvolution from the vocal-tract filter is posed as a SBD problem. We employ ALPA for SBD and show that excitation obtained is sparser than the excitations obtained using sparse linear prediction, smoothed `1=`2 sparse blind deconvolution algorithm, and majorization-minimization-based sparse deconvolution techniques. We also consider the problem of epoch estimation and show that epochs estimated by ALPA in both clean and noisy conditions are closer to the instants indicated by the electroglottograph when with to the estimates provided by the zero-frequency ltering technique, which is the state-of-the-art epoch estimation technique. In the third part, we consider the problem of deconvolution of a specific class of continuous-time signals called nite-rate-of-innovation (FRI) signals, which are not bandlimited, but specified by a nite number of parameters over an observation interval. The signal is assumed to be a linear combination of delayed versions of a prototypical pulse. The reconstruction problem is posed as a 2-D SBD problem. The kernel is assumed to have a known form but with unknown parameters. Given the sampled version of the FRI signal, the delays quantized to the nearest point on the sampling grid are rst estimated using proximal-operator-based alternating `p `2 algorithm (ALPAprox), and then super-resolved to obtain o -grid (O. G.) estimates using gradient-descent optimization. The overall technique is termed OG-ALPAprox. We show application of OG-ALPAprox to a particular modality of super-resolution microscopy (SRM), called stochastic optical reconstruction microscopy (STORM). The resolution of the traditional optical microscope is limited by di raction and is termed as Abbe's limit. The goal of SRM is to engineer the optical imaging system to resolve structures in specimens, such as proteins, whose dimensions are smaller than the di raction limit. The specimen to be imaged is tagged or labeled with light-emitting or uorescent chemical compounds called uorophores. These compounds speci cally bind to proteins and exhibit uorescence upon excitation. The uorophores are assumed to be point sources and the light emitted by them undergo spreading due to di raction. STORM employs a sequential approach, wherein each step only a few uorophores are randomly excited and the image is captured by a sensor array. The obtained image is di raction-limited, however, the separation between the uorophores allows for localizing the point sources with high precision. The localization is performed using Gaussian peak- tting. This process of random excitation coupled with localization is performed sequentially and subsequently consolidated to obtain a high-resolution image. We pose the localization as a SBD problem and employ OG-ALPAprox to estimate the locations. We also report comparisons with the de facto standard Gaussian peak- tting algorithm and show that the statistical performance is superior. Experimental results on real data show that the reconstruction quality is on par with the Gaussian peak- tting.

La progressivité du claquage des oxydes de grille d'épaisseurs inférieures à 20 nm permet d'envisager une prolongation de la durée de vie des circuits. Cet enjeu majeur de la fiabilité contemporaine requiert des modèles adaptés afin de contrôler la variabilité des paramètres induites par le claquage. Après avoir étudié l'impact d'une fuite de courant sur une couche chargée, nous avons mis au point un modèle bas niveau de simulation par éléments finis, capable de reproduire la dérive des paramètres mesurée sur des dispositifs du nœud 45 nm. Des lois empiriques de ces dérives ont été injectées dans un modèle compact du transistor dégradé, simplifié par nos observations originales de la dépolarisation du canal et de la répartition des courants. Finalement nous avons simulé l'impact du claquage sur le fonctionnement de circuits simples et estimés la dérive de leurs paramètres tels que l'augmentation de la consommation due au claquage. / Breakdown (BD) progressivity for oxides thicker than 20nm may allow circuit lifetime extension; for design purpose and reliability questions, it is now very important to include soft BD failure in compact models in order to predict circuit's parameters variability. After studying the impact of current leakage on a charged layer, we set up a low level simulation model, able to reproduce parameters deviation measured on MOSFET from the 45nm node. Empirical laws of parameter's variability due to this degradation have been used to build up a compact model of damaged device. Our observations have allowed several improvements of BD understanding and led to major simplifications in BD compact modelling. Our simulations of small circuits show a good agreement with published measures and allow an estimation of BD impact on circuits, such as circuit's parameters deviation and power consumption increase estimation.

This thesis deals with the topic of  model-based design and application of simulation in system design. In the introduction, the thesis discusses the history of software development process and outlines the current status. The aim is to demonstrate a model-driven design on a case study of conference system. There will be presented formalisms of DEVS and OOPN  together with experimental tools PNtalk and SmallDEVS that allow to work with these formalisms. The resulting model of conference system will be deployed as part of a web application using a framework Seaside in the Squeak environment.