Ubiquitylation is an indispensable post-translational modification system in eukaryotic cells that leads to the covalent attachment of a small ubiquitin (Ub) protein onto a target. The traditional and best-characterized role for ubiquitylation is a fundamental regulatory mechanism whereby target proteins are tagged with a characteristic Lys48-linked Ub chain that signals for their elimination through proteasomal degradation. Challenging this conventional wisdom is the finding that some ubiquitylated proteins are modified by Ub chains linked through Lys63, providing a molecular signal that is thought to be structurally and functionally distinct from Lys48-linked Ub chains. Of further interest and significance is that the Lys63-linked Ub chains are apparently synthesized through a novel biochemical mechanism employing a unique complex formed between a true Ub conjugating enzyme (E2), Ubc13, and an E2-variant (Uev), Mms2 (or Uev1A). The goal of this thesis was to employ structural and functional approaches in order to better characterize the Ubc13-Mms2 Lys63-linked Ub conjugation complex. <p>Error-free DNA damage tolerance (DDT) in the budding yeast is dependent on Lys63-linked Ub chains synthesized by Ubc13-Mms2 and thus provided the opportunity to experimentally test the function of the human UBC13 and MMS2 genes in a simple model organism. Human UBC13 and MMS2 were each shown to function in place of their yeast counterparts and in accordance, human Ubc13 was shown to physically interact with yeast Mms2, and vice versa. Two human MMS2 homologs were also tested and it was determined that UEV1A but not UEV1B can function in place of mms2 in yeast DDT. Physical interactions were observed between Ubc13 and Uev1A, but not between Ubc13 and Uev1B, suggesting that Ubc13-Uev complex formation is required for function. <p>In collaboration with a research group at the University of Alberta, crystal structure and NMR data were used to develop a mechanistic model for the conjugation of Lys63-linked Ub chains by the Ubc13-Mms2 heterodimer, whereby the special orientation of two Ub molecules facilitates a specific Ub-Ub linkage via Lys63. In order to help support the in vitro model and to determine how the Ubc13-Mms2 structure relates to biological function, I used a structure-based approach to direct the creation of point mutations within four key regions of the Ubc13-Mms2 heterodimer; the Ubc13 active-site, the Ubc13-E3 (Ub ligating enzyme) interface, the Mms2-Ub interface, and the Ubc13-Mms2 interface. <p>Underscoring the importance of the Ub conjugation by Ubc13-Mms2, a Ubc13-C87S active-site mutation was created that could bind to Mms2 but was unable to function in DDT. Regarding the Ubc13-E3 interface, a single Ubc13-M64A point mutation had a potent effect on disrupting Ubc13 function in DDT, as well as its physical interaction with Rad5, TRAF6, and CHFR. The results suggest that different RING finger E3s use the same Ubc13 surface to sequester the Ub conjugation activity of Ubc13-Mms2. Two human Mms2 mutations at Ser32 and Ile62, which are contained within the Mms2-Ub interface, were found to reduce the ability of Mms2 to bind Ub. When the corresponding yeast mutations are combined, a synergistic loss in DDT function is observed. The relative orientation of Ser32 and Ile62 suggests that the Mms2 and Tsg101 Uev families use different Uev surfaces to physically interact with Ub. A 200 ìM dissociation constant for the wild-type Mms2-Ub interaction was also determined. The systematic mutagenesis and testing of 14 Ubc13-Mms2 interface residues led to mutants with partial or complete disruption of binding and function. Using this data, a model involving the insertion of a specific Mms2-Phe residue into a unique Ubc13 hydrophobic pocket was created to explain the specificity of Mms2 for Ubc13, and not other E2s. In addition, the dissociation constant for the wild-type Ubc13-Mms2 heterodimer was determined to be approximately 50 nM. <p>The structural and functional studies strongly support the notion that Ubc13-Mms2 complex has the unique ability to conjugate Lys63-linked Ub chains. However, several reported instances of Lys63-linked Ub chains in vivo have not yet been attributed to Ubc13 or Mms2. To address the disparity I was able to demonstrate and map a physical interaction between Mms2 and Rsp5, an E3 implicated in Lys63-linked Ub conjugation. Surprisingly, it was found that MMS2 is not responsible for the RSP5-dependent Lys63-linked Ub conjugation of a plasma membrane protein. A possible explanation for the apparent paradox is presented.

Ubiquitylation is an indispensable post-translational modification system in eukaryotic cells that leads to the covalent attachment of a small ubiquitin (Ub) protein onto a target. The traditional and best-characterized role for ubiquitylation is a fundamental regulatory mechanism whereby target proteins are tagged with a characteristic Lys48-linked Ub chain that signals for their elimination through proteasomal degradation. Challenging this conventional wisdom is the finding that some ubiquitylated proteins are modified by Ub chains linked through Lys63, providing a molecular signal that is thought to be structurally and functionally distinct from Lys48-linked Ub chains. Of further interest and significance is that the Lys63-linked Ub chains are apparently synthesized through a novel biochemical mechanism employing a unique complex formed between a true Ub conjugating enzyme (E2), Ubc13, and an E2-variant (Uev), Mms2 (or Uev1A). The goal of this thesis was to employ structural and functional approaches in order to better characterize the Ubc13-Mms2 Lys63-linked Ub conjugation complex. <p>Error-free DNA damage tolerance (DDT) in the budding yeast is dependent on Lys63-linked Ub chains synthesized by Ubc13-Mms2 and thus provided the opportunity to experimentally test the function of the human UBC13 and MMS2 genes in a simple model organism. Human UBC13 and MMS2 were each shown to function in place of their yeast counterparts and in accordance, human Ubc13 was shown to physically interact with yeast Mms2, and vice versa. Two human MMS2 homologs were also tested and it was determined that UEV1A but not UEV1B can function in place of mms2 in yeast DDT. Physical interactions were observed between Ubc13 and Uev1A, but not between Ubc13 and Uev1B, suggesting that Ubc13-Uev complex formation is required for function. <p>In collaboration with a research group at the University of Alberta, crystal structure and NMR data were used to develop a mechanistic model for the conjugation of Lys63-linked Ub chains by the Ubc13-Mms2 heterodimer, whereby the special orientation of two Ub molecules facilitates a specific Ub-Ub linkage via Lys63. In order to help support the in vitro model and to determine how the Ubc13-Mms2 structure relates to biological function, I used a structure-based approach to direct the creation of point mutations within four key regions of the Ubc13-Mms2 heterodimer; the Ubc13 active-site, the Ubc13-E3 (Ub ligating enzyme) interface, the Mms2-Ub interface, and the Ubc13-Mms2 interface. <p>Underscoring the importance of the Ub conjugation by Ubc13-Mms2, a Ubc13-C87S active-site mutation was created that could bind to Mms2 but was unable to function in DDT. Regarding the Ubc13-E3 interface, a single Ubc13-M64A point mutation had a potent effect on disrupting Ubc13 function in DDT, as well as its physical interaction with Rad5, TRAF6, and CHFR. The results suggest that different RING finger E3s use the same Ubc13 surface to sequester the Ub conjugation activity of Ubc13-Mms2. Two human Mms2 mutations at Ser32 and Ile62, which are contained within the Mms2-Ub interface, were found to reduce the ability of Mms2 to bind Ub. When the corresponding yeast mutations are combined, a synergistic loss in DDT function is observed. The relative orientation of Ser32 and Ile62 suggests that the Mms2 and Tsg101 Uev families use different Uev surfaces to physically interact with Ub. A 200 ìM dissociation constant for the wild-type Mms2-Ub interaction was also determined. The systematic mutagenesis and testing of 14 Ubc13-Mms2 interface residues led to mutants with partial or complete disruption of binding and function. Using this data, a model involving the insertion of a specific Mms2-Phe residue into a unique Ubc13 hydrophobic pocket was created to explain the specificity of Mms2 for Ubc13, and not other E2s. In addition, the dissociation constant for the wild-type Ubc13-Mms2 heterodimer was determined to be approximately 50 nM. <p>The structural and functional studies strongly support the notion that Ubc13-Mms2 complex has the unique ability to conjugate Lys63-linked Ub chains. However, several reported instances of Lys63-linked Ub chains in vivo have not yet been attributed to Ubc13 or Mms2. To address the disparity I was able to demonstrate and map a physical interaction between Mms2 and Rsp5, an E3 implicated in Lys63-linked Ub conjugation. Surprisingly, it was found that MMS2 is not responsible for the RSP5-dependent Lys63-linked Ub conjugation of a plasma membrane protein. A possible explanation for the apparent paradox is presented.

The first part of the thesis describes a systematic function analysis that identified in vitro E2 partners for ten different HECT E3 ligase proteins. Using mass spectrometry, the linkage composition for the resulting autoubiquitylation products of a number of functional E2-HECT pairs was determined. HECT domains from different subfamilies catalyze the formation of very different types of Ub chains, largely independent of the E2 in the reaction. The second part of the thesis describes the characterization of the RAD6-interactome. Using affinity purification coupled with mass spectrometry, I identified a novel RAD6-interacting E3 ligase, KCMF1, which binds to a different surface on RAD6 than the other RAD6-associated E3 ligases. KCMF1 also recruits additional proteins to RAD6, and this new complex points to novel RAD6 functions. Interestingly, the RAD6A R11Q mutant polypeptide, found in X-linked mental retardation patients specifically loses the interaction with KCMF1, but not with other RAD6-associated E3 ligases.

The first part of the thesis describes a systematic function analysis that identified in vitro E2 partners for ten different HECT E3 ligase proteins. Using mass spectrometry, the linkage composition for the resulting autoubiquitylation products of a number of functional E2-HECT pairs was determined. HECT domains from different subfamilies catalyze the formation of very different types of Ub chains, largely independent of the E2 in the reaction. The second part of the thesis describes the characterization of the RAD6-interactome. Using affinity purification coupled with mass spectrometry, I identified a novel RAD6-interacting E3 ligase, KCMF1, which binds to a different surface on RAD6 than the other RAD6-associated E3 ligases. KCMF1 also recruits additional proteins to RAD6, and this new complex points to novel RAD6 functions. Interestingly, the RAD6A R11Q mutant polypeptide, found in X-linked mental retardation patients specifically loses the interaction with KCMF1, but not with other RAD6-associated E3 ligases.