Thesis (Ph. D.)--University of Washington, 2000. / Vita. Includes bibliographical references (leaves 74-91).

Changes in cellular redox status are implicated in the regulation of developmental and defence-related responses. The absence of S-nitrosoglutathione reductase (GSNOR) function in Arabidopsis leads to an accumulation of cellular S-nitrosoglutathione (GSNO), a mobile reservoir of nitric oxide (NO) which impacts the cellular redox tone. Consequently, the GSNOR knockout mutant, atgsnor1-3 displays defects in growth, time to flowering and pathogen resistance. Although it is now well established that GSNO is a key redox signalling molecule, the molecular mechanisms that underpin GSNO function remains largely unknown. RAP1 (REDOX-ASSOCIATED PROTEIN 1) was identified based on its dynamic changes of expression in atgsnor1-3 and sid2 plants upon avirulent Pseudomonas syringae pv. tomato (Pst) DC3000 (avrB) challenge. Pathogen-induced RAP1 expression was shown to be independent of the plant hormones salicylic acid, jasmonic acid, abscisic acid and ethylene. Recombinant RAP1 protein was shown to exhibit E3 ligase activity in vitro. Application of the NO donors (GSNO and Cysteine-NO (CysNO)) reduced the E3 ligase activity of RAP1 significantly. Biotinswitch analysis showed that RAP1 was S-nitrosylated and site-directed mutagenesis of RAP1 suggested that the S-nitrosylated site is the cysteine residue C325. The rap1 line does not show obvious developmental phenotypes, however, overexpressing RAP1 enhanced lateral root branching in young seedlings. Overexpression of a truncated RAP1 (RAP1ΔRING) led to a loss of apical dominance. In addition, rap1/rap2 double mutants showed delayed flowering, suggesting RAP1 might be involved in the regulation of plant growth and development. RAP1 may also be involved in plant defence, as rap1, rap2 and rap1/rap2 mutants exhibited increased susceptibility to PstDC3000 and Arabidopsis powdery mildew. Interestingly, rap1 plants showed enhanced resistance to methyl viologen (MV), which is in line with the phenotype of atgsnor mutants. Also, expression of RAP1 was rapidly inducible by ultraviolet-B (UV-B) light. As RAP1 expression and RAP1 E3 ligase activity are redox-related, it is speculated that RAP1 may be involved in redoxmediated regulation of a broad range of physiological responses.

The BCA2 E3 ligase is expressed in a majority of invasive breast cancers. BCA2 has inherent autoubiquitination activity which contributes to cell migration and proliferation processes. Here, ten novel BCA2 binding proteins were found using yeast and bacterial screening. Two of which were human homolog of Rad23 variant A (hHR23a) and 14-3-3σ. In vivo and in vitro assays confirmed that both hHR23a and 14-3-3σ bound BCA2 and were co-expressed with BCA2 in breast cancer cells. Interaction of BCA2 with hHR23a and 14-3-3σ affect the autoubiquitination and auto-degradation activity of BCA2. Multi-ubiquitination of hHR23a-bound BCA2 was dramatically lower than that of free BCA2, this corresponded to increased BCA2 expression and half-life. Furthermore, phosphorylated BCA2 protein was stabilized by interaction with 14-3-3σ, via substrate inhibition of BCA2 autoubiquitination. High expression of BCA2 is correlated with grade in breast cancer and regulation of this E3 ligase’s activity may be important to cancer progression.

The BCA2 E3 ligase is expressed in a majority of invasive breast cancers. BCA2 has inherent autoubiquitination activity which contributes to cell migration and proliferation processes. Here, ten novel BCA2 binding proteins were found using yeast and bacterial screening. Two of which were human homolog of Rad23 variant A (hHR23a) and 14-3-3σ. In vivo and in vitro assays confirmed that both hHR23a and 14-3-3σ bound BCA2 and were co-expressed with BCA2 in breast cancer cells. Interaction of BCA2 with hHR23a and 14-3-3σ affect the autoubiquitination and auto-degradation activity of BCA2. Multi-ubiquitination of hHR23a-bound BCA2 was dramatically lower than that of free BCA2, this corresponded to increased BCA2 expression and half-life. Furthermore, phosphorylated BCA2 protein was stabilized by interaction with 14-3-3σ, via substrate inhibition of BCA2 autoubiquitination. High expression of BCA2 is correlated with grade in breast cancer and regulation of this E3 ligase’s activity may be important to cancer progression.

Würzburg, Universiẗat, Diss., 2005.

Loss-of-function genetic screens are a powerful approach to identify the genes involved in biological processes. For nearly a century, forward genetic screens in model organisms have provided enormous insight into many cellular processes. However, the difficulty in generating and recovering bi-allelic mutations in diploid cells severely hindered the performance of forward genetic screens in mammalian cells. The development of a retroviral gene-trap vector to mutagenise the human near-haploid KBM7 cell line transformed forward genetic screens in human cells. The re-purposing of the microbial CRISPR/Cas9 system now offers an effective method to generate gene knockouts in diploid cells. Here, I performed a head-to-head comparison of retroviral gene-trap mutagenesis screens and genome-wide CRISPR knockout screens in KBM7 cells. The two screening approaches were equally effective at identifying genes required for the endoplasmic reticulum (ER)-associated degradation of MHC class I molecules. The ER-resident enzyme HMG-CoA reductase (HMGCR) catalyses the rate-limiting step in the cholesterol biosynthesis pathway and is targeted therapeutically by statins. To maintain cholesterol homeostasis, the expression of HMGCR is tightly regulated by sterols transcriptionally and post-translationally. Sterols induce the association of HMGCR with Insig proteins, which recruit E3 ubiquitin ligase complexes to mediate degradation of HMGCR by the ubiquitin proteasome system. However, the identity of the E3 ligase(s) responsible for HMGCR ubiquitination is controversial. Here, I use a series of genome-wide CRISPR knockout screens using a fluorescently-tagged HMGCR exogenous reporter and an endogenous HMGCR knock-in as an unbiased approach to identify the E3 ligases and any additional components required for HMGCR degradation. The CRISPR screens identified a role for the poorly characterised ERAD E3 ligase RNF145. I found RNF145 to be functionally redundant with gp78, an E3 ligase previously implicated in HMGCR degradation, and the loss of both E3 ligases was required to significantly inhibit the sterol-induced degradation and ubiquitination of HMGCR. A focused E3 ligase CRISPR screen revealed that the combined loss of gp78, RNF145 and Hrd1 was required to completely block the sterol-induced degradation of HMGCR. I present a model to account for this apparent complexity.

The ubiquitination of proteins controls a multitude of physiological processes. This versatility of ubiquitin as a molecular signal arises from the diverse ways by which it can be attached to target proteins. Different ubiquitination patterns are then translated into different downstream consequences. Due to the enormous complexity of possible ubiquitin modifications, the ubiquitination machinery must be highly specific and tightly controlled. Ubiquitination proceeds through an enzymatic cascade, the last step of which is catalyzed by the E3 enzyme family. E3 enzymes are the crucial regulators since they dictate the specificity of substrate selection and modification. Deregulation of the HECT-type ubiquitin ligase E6AP (UBE3A) is implicated in human papilloma virus-induced cervical tumorigenesis and several neurodevelopmental disorders. Yet the structural underpinnings of activity, regulation and specificity in this crucial ligase are incompletely understood. One aim of this study was to unravel the role of the a1’-helix N-terminal to the HECT domain that was found to be a key element mediating regulation and oligomerization in other HECT ligases. I found that most N-terminally extended HECT domain constructs were insoluble when expressed in E. coli, indicating that additional regions N-terminal to the tested fragments may be essential to protect this highly hydrophobic helix from causing aggregation. Another question addressed in this study was how E6AP builds ubiquitin chains. Using single-turnover experiments, I showed that ubiquitin-loaded E6AP is unable to transfer an additional ubiquitin molecule onto a stably linked ubiquitin-E6AP complex. This indicates that E6AP cannot assemble chains on its active site and may instead follow a sequential addition mechanism in which one ubiquitin molecule is transferred at a time to the target protein. Using NMR spectroscopy and extensive mutational analyses, the determinants of ubiquitin recognition by the C-lobe of E6AP were unraveled and assigned to particular steps in the catalytic cycle. A functionally critical interface was identified that is specifically required during thioester formation between the C-terminus of ubiquitin and the ligase active site. This interface resembles the one utilized by NEDD4-type enzymes, suggesting a conserved ubiquitin binding mode across HECT ligases, independent of their linkage specificities. Moreover, I identified critical surface patches on ubiquitin and in the N- and C-terminal portions of the catalytic domain of E6AP that are important for the subsequent step of isopeptide bond formation. I also uncovered key determinants of the Lys48-linkage specificity of E6AP, both in the E6AP HECT domain and ubiquitin itself. This includes the C-terminal tail of E6AP and a hydrophilic surface region of ubiquitin in proximity to the acceptor site, Lys48. It is thus tempting to speculate that ubiquitin linkage formation by E6AP is substrate-assisted. Taken together, my results improve our mechanistic understanding of the structure-function relationship between E6AP and ubiquitin, thus providing a basis for ultimately manipulating the functions of this HECT ligase for therapeutic applications. / Die Ubiquitinierung von Proteinen ist an nahezu jedem physiologischen Prozess beteiligt. Die Vielseitigkeit mit der Ubiquitin als molekulares Signal fungiert, rührt von den vielfältigen Möglichkeiten her, wie es an Zielproteine gebunden werden kann. Verschiedene Ubiquitinierungsmuster rufen unterschiedliche biologische Ereignisse hervor. Angesichts der enormen Komplexität möglicher Ubiquitinierungsmodifikationen muss die Ubiquitinierungs-maschinerie hochspezifisch und streng kontrolliert sein. Die Ubiquitinierung erfolgt über eine enzymatische Kaskade. Der letzte Schritt wird hierbei durch die Enzymfamilie der Ubiquitin-Ligasen katalysiert. Ubiquitin-Ligasen sind primär für die Spezifität in Substraterkennung und Ubiquitin-Kettenbildung verantwortlich. Misregulation der HECT-Ligase E6AP fördert die durch humane Papillomaviren induzierte Tumorentwicklung im Gebärmutterhals und ist mit zwei schweren neurologischen Krankheiten verbunden. Strukturelle Einzelheiten über den Mechanismus, die Regulation und die Spezifität dieser wichtigen Ligase sind jedoch weitgehend unbekannt. Für verschiedene HECT-Ligasen wurde gezeigt, dass die a1‘-Helix N-terminal zur HECT-Domäne ein Schlüsselelement für die Regulation und den Oligomerisierungszustand der Enzyme darstellt. In dieser Arbeit konnte gezeigt werden, dass die Helix eine wichtige Funktion für die Stabilität von E6AP erfüllt. Der Großteil N-terminal verlängerter, in E. coli exprimierter HECT-Domänen-Konstrukte war unlöslich, was darauf hindeutet, dass N-terminal gelegene Regionen hydrophobe Bereiche des Proteins vor Aggregation schützen. Eine weitere Fragestellung dieser Arbeit befasste sich mit dem Mechanismus der Ubiquitin-Kettenbildung durch E6AP. Mit ‘single-turnover‘-Experimenten konnte gezeigt werden, dass ein über einen Thioester gebundenes Ubiquitin von E6AP nicht auf einen stabil verknüpften Ubiquitin-E6AP-Komplex übertragen werden kann. Dies deutet daraufhin, dass E6AP keine Ketten auf dem katalytischen Cystein aufbauen kann und stattdessen einem sequentiellen Additionsmechanismus der Ubiquitin-Kettenbildung folgt. Mithilfe von NMR Spektroskopie und umfangreicher Mutagenese-Studien wurde eine Interaktion zwischen dem C-Lobe von E6AP und Ubiquitin gefunden, die während der Thioesterbildung zwischen dem C-Terminus von Ubiquitin und dem aktiven Zentrum von E6AP gebraucht wird. Diese Interaktionsfläche ähnelt derer der NEDD4-Familie, was auf einen konservierten Bindungsmodus der HECT-Ligasen an Ubiquitin im ersten Reaktionsschritt hindeutet, ungeachtet der jeweiligen Kettenspezifitäten. Verschiedene Oberflächen auf Ubiquitin und E6AP, sowohl auf dem C-Lobe als auch auf dem N-Lobe, konnten identifiziert werden, die für die Bildung einer Isopeptidbindung zwischen zwei Ubiquitin-Molekülen von Bedeutung sind. Neben dem C-Terminus von E6AP wurde eine hydrophile Oberfläche auf Ubiquitin in unmittelbarer Nähe zum Akzeptor Lys48 gefunden, die wichtig für die Lys48-spezifische Ubiquitin-Kettenbildung ist. Der Gedanke liegt nahe, dass die Ubiquitin-Kettenbildung durch E6AP über Substratunterstützte Katalyse verläuft. Zusammenfassend erweitern diese Ergebnisse maßgeblich unser Verständnis der Erkennung von Ubiquitin durch die HECT-Ligase E6AP und können möglicherweise dazu beitragen Wirkstoffe zu entwickeln, welche eine Fehlregulierung von E6AP ausgleichen können.

Maintenance of genomic integrity is dependent on the duplication of chromosomes, only once per cell cycle. Highly conserved mechanisms for the regulation of chromosome replication exists to ensure that the genome is copied only once. The Cdc45-MCM-GINS (CMG) DNA helicase which is the core of the eukaryotic replication complex, has been shown to be extensively regulated by post translational modifications, during its assembly. Therefore, it is not inconceivable that the process to unload the replication complex would also be a conserved and regulated process. In 2014, our lab discovered that the CMG complex undergoes post-translational modification in the form of ubiquitylation on one of the subunits of CMG, leading to its disassembly from the chromatin. Though the main players in the disassembly of CMG were known, viz the E3 ligase SCFDia2 and segregase Cdc48, very little was known about the mechanism of CMG disassembly. In the process of learning more about the disassembly of the replicative helicase from chromatin, I reconstituted the ubiquitylation of CMG and thereafter the disassembly of CMG helicase in vitro. My work resulting in the reconstitution of CMG disassembly in vitro is the first example of the disassembly of a multi-subunit physiological substrate of Cdc48. Though CMG is ubiquitylated in yeast extracts in vitro, it does not lead to its disassembly and therefore led me to find conditions necessary for the efficient ubiquitylation of CMG. I have further shown that purifying the E3 ligase associated CMG can be efficiently ubiquitylated in a semi-reconstituted system consisting of purified factors, necessary for the ubiquitylation of substrate. I investigated whether this efficiently ubiquitylated CMG can be disassembled by purified Cdc48 and associated co-factor Ufd1/Npl4 in vitro and found that disassembly is dependent on K48 linked poly-ubiquitylation of CMG. I have found that the reconstituted poly-ubiquitylation of CMG is restricted to the Mcm7 subunit of CMG, recapitulating the ubiquitylation of CMG in vivo, and my data points out that there are multiple sites of ubiquitylation on Mcm7. Through this work, I have also found that ubiquitylated Mcm7 no longer associates with the rest of the CMG components after disassembly of CMG. My assays and findings, open the door towards dissecting the molecular mechanism of the disassembly of CMG in greater detail.

Estudos preliminares através de microarray nos mostraram que a E3 ligase Mdm2 foi regulado positivamente no músculo de animais hipertireoideos. Dessa forma, nós inferimos uma possível relação de Mdm2 com a atrofia causada por T3. Para testar nossa hipótese, ratos foram induzidos ao hipertireoidismo para análises subsequentes. Concomitante com a perda de massa muscular foi confirmado um aumento da expressão de Mdm2 tanto no nível gênico (p<0.05) quanto protéico. Interessantemente, Mdm2 foi preferencialmente expresso em fibras tipo I, mostrando maior sensibilidade dessas fibras ao T3. Além disso, foi observado uma diminuição severa na expressão de Pax7/MyoD associado à superexpressão de Mdm2, sugerindo inatividade das células satélites. Surpreendentemente, a inibição de Mdm2 em miotubos cultivados provocou uma diminuição severa no diâmetro destes (~35%, p<0.05), ou seja, tal inibição foi incapaz de minimizar a proteólise muscular causada por T3. Portanto, nós concluímos que a responsividade de Mdm2 ao T3 agiria como um mecanismo compensatório numa tentativa de minimizar a proteólise muscular causada pelo hipertireoidismo. Esta conclusão é reforçada pela atrofia observada em miotubos durante a inibição de Mdm2 sem a presença de T3. / Previous studies in our lab through microarray assay observed Mdm2, an E3 ligase, up regulated in soleus muscle from hyperthyroid rats. In this sense, we inferred that Mdm2 could be related to muscle atrophy caused by T3. To test our hypothesis, rats were induced to experimental hyperthyroidism for subsequent analysis. Along the muscle mass loss, the increase on Mdm2 gene expression was confirmed (p<0.05) as well as protein expression by RT-PCR and Western Blot, respectively. Interestingly, Mdm2 was expressed predominantly in fiber I type during T3 treatment, demonstrating a higher sensibility when compared to type II fiber. Moreover, it was observed a severe decrease in Pax7/MyoD labeling, associated to an increase on Mdm2 labeling, suggesting that T3 could be associated with inactivation of satellite cells. Surprisingly, Mdm2 inhibition in myotubes have induced severe decrease on myotubes diameter (~35%, p<0.05), in other words, Mdm2 inhibition was not able to decrease muscle proteolysis during high levels of T3. Thus, the increase on Mdm2 levels could be a compensatory effect to reduce the muscle mass loss during T3 treatment. This conclusion is highlighted by the myotubes atrophy observed during the Mdm2 inhibition without T3 treatment.

Le blé tendre, Triticum aestivum, représente une grande ressource dans l’alimentation humaine mais également dans l’industrie. En conséquence, la taille finale du grain de blé constitue une cible privilégiée des programmes de sélections variétales. Pour ces raisons, comprendre les mécanismes moléculaires qui contrôlent le développement du grain, en particulier lors des phases précoces de la mise en place des structures cellulaires et leur remplissage par des réserves, constitue un enjeu majeur. Le développement du grain de blé est un processus complexe qui nécessite l’intervention séquentielle ou combinée d’un très grand nombre de gènes et de voies métaboliques. Parmi ces voies, le métabolisme carboné (en particulier celui du saccharose), les voies de signalisation par les hormones et la voie Ubiquitine / Protéasome 26S (UPS) semblent jouer un rôle déterminant dans la taille finale et donc le rendement en grain chez les céréales. Pour étudier le développement du grain de blé tendre, des plantes de la variété Récital ont été cultivées en serre dans des conditions optimales sans contraintes. Les grains ont été récoltés à onze stades de développement après floraison, allant de 40°CJ (soit deux jours après floraison) à 500°CJ (soit 25 jours après floraison). Les ARN totaux ont été extraits à partir de ces grains et utilisés pour analyser l’expression des gènes par une approche transcriptomique, soit en utilisant une lame « dédiée », soit en utilisant une lame Nimblegen comprenant 39 179 gènes. Une analyse différentielle utilisant le test LIMMA a permis d’identifier 9284 gènes différentiellement exprimés. L’analyse globale de ces gènes a montré que des modifications transcriptionnelles majeures ont lieu entre les stades 80 et 120 °CJ ainsi qu’entre 220 et 240°CJ. La répartition de ces 9284 en 10 clusters, en fonction de leurs profils d’expression, permet d’identifier les gènes activés en début de la phase de division cellulaire chez le grain, ceux activés pendant la phase de remplissage et ceux présentant un profil dit en « cloche ». Parmi les gènes différentiellement exprimés, nous nous sommes intéressés à ceux qui codent pour des E3 ligases impliquées dans la voie UPS et aux gènes relatifs à 7 hormones végétales (auxine, acide abscissique, acide jasmonique, brassinostéroïdes, cytokinines, gibbérellines et éthylène). Nous avons alors identifié 173 gènes codant pour des E3 ligases (dont certaines sont également des récepteurs hormonaux) et 126 gènes impliqués dans les voies hormonales. Un modèle global décrivant la chronologie d’intervention de ces gènes a été proposé. La majorité des gènes E3 de type SCF (SKP1-Cullin-Fbox), APC/C, Cul3-BTB et Ubox interviendrait dans les phases précoces du développement du grain de blé. Parallèlement, la majorité des gènes relatifs à l’auxine, à l’acide jasmonique et aux brassinostéroïdes interviendrait lors des phases de divisions cellulaires alors que les gènes relatifs à l’éthylène et à l’acide abscissique interviendraient dans les phases de remplissage. Par ailleurs, une méta-analyse a été réalisée et a permis d’identifier 26 gènes candidats codants pour des E3 ligases ainsi que de 12 gènes candidats impliqués dans les voies hormonales qui seraient préférentiellement exprimés dans l’albumen, un tissu du grain à haute valeur agro-économique. Le modèle proposé et l’identification de ces gènes candidats établissent un cadre pour de futures études visant à comprendre les mécanismes moléculaires contrôlant le développement du grain de blé. / Wheat grain is an important source of food, feed, and industrial raw materials, but current production levels cannot meet world needs. Elucidation of the molecular mechanisms underlying wheat grain development will contribute valuable information to improving wheat cultivation. One of the most important mechanisms implicated in plant developmental processes is the Ubiquitin-Proteasome System (UPS). Among several implications of the UPS, it has become clear that it plays an essential role in hormone signaling. In particular E3 ubiquitin ligases, from the UPS, have been demonstrated to play critical roles in hormone perception and signal transduction. During these work, wheat cv. Recital were grown in optimum growth conditions. By comparing eleven consecutive time-points from 40°CJ (2 days after anthesis) to 500°CJ (around 25 days after anthesis), 9284 differentially expressed genes were identified during this study. A comparison of these genes in terms of time revealed dynamic transcript accumulation profiles with major re-programming events that occurred during the time intervals of 80-120°Cdays and 220-240°Cdays. The gene expression comparison allows observing genes potentially involved in cell division or grain filling stage. An emphasis was made on the E3 ligases and hormone-related genes (Abscisic acid, Auxin, Brassinosteroid, Cytokinine, Gibberellic acid, Ethylene and Jasmonic acid). 173 E3 ligase coding genes and 126 hormone–related genes were found to be differentially expressed during the cell division and grain filling stages, with a different expression profile for each family. A model describing the timing of the involvement of these genes is proposed to provide a framework for the design of future experiments and for the identification of genes and pathways for further characterization. A majority of the E3 SCF (SKP1-Cullin-F-box), APC/C, Cul3-BTB and Ubox are found expressed in early wheat developmental stages (cell division stage). A majority of auxin, jasmonic acid and brassinostéroïde related genes were found to be up-regulated in early wheat developmental stages while ethylene and abscisic acid related genes were found to be activated during grain filling stage. The differential expression of genes involved in E3 ligase pathways and plant hormone signalling suggested that phytohormones and UPS crosstalk might play a critical role in the wheat grain developmental process. A meta-analysis of these genes led to the identification of 26 E3 ligase candidate genes and 12 hormones-related candidate genes that are preferentially expressed in the endosperm. The functional model that we proposed and the identification of candidate genes should help to better understand wheat grain development.